Speak at The Fifth Elephant 2026 Annual Conference
Share you work with the community
Accepting submissions
Not accepting submissions
Resources:
- How to create a strong proposal - by Vivek Pemavat, editor of Rootconf Platform Engineering meet-ups: https://hasgeek.com/rootconf/call-for-submissions-platform-engineering-meet-ups/sub/rootconf-platform-engineering-talk-guidelines-MWUtatELnYgQb3dLeU2FA6
|
ss
sooraj shankar Memory Is a Data System: State, Search, and History for AI AgentsAI agent memory is often discussed as a prompting problem or a vector search problem. In production, it behaves much more like a data system. Agents need to store facts, preferences, decisions, intermediate notes, and shared organizational context. That memory has to be named, scoped, searched, updated, versioned, inspected, and eventually corrected. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
VS
Vivek Sinha High Concurrency & Low Latency Serving on Apache IcebergDescription Everyone is putting their data into Apache Iceberg; almost no one is serving sub-second queries directly from it. Once data lands in Iceberg, a familiar question arises: how do you power real-time experiences without duplicating it in yet another serving system? This challenge is especially sharp in observability workloads like RUM, clickstream, and APM, and in customer-facing analyti… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
Siddhant Agarwal Your AI Agent is lying to you: Observability for LLM systems in productionYou have shipped your LLM-powered agent. Congrats. Now do you actually know what it is doing? Most teams flying blind in production only discover issues when users complain, by which point the damage is already done. This talk dives deep into the observability gap in GenAI systems: why traditional APM tools were never designed for non-deterministic, multi-step agentic workflows, and why bolting t… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
KS
Karthik Shashidhar How to drive adoption and derive value from your AI-for-data agentDescribe your session in 2 paragraphs Companies are struggling to get value from AI-for-data agents. They have invested considerable time and energy in onboarding such systems but found usage and adoption lacking. The reasons are manifold - business users lack confidence in the systems, providing the right context is a challenge, people simply don’t know what questions to ask. So rather than usin… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
kg
krishan goyal Optimizing Data Ingestion in Apache PinotOptimizing Data Ingestion in Apache Pinot The Problem more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
Pramod Biligiri Editor, The Fifth Elephant 2024 & 2025 editions JK
Jagadish K Editor Session formats at The Fifth Elephant: Birds of Feather sessionsBirds of a Feather (BOF) sessions A Birds of a Feather session is a focused, practitioner-led conversation on a specific topic. It is not a talk, not a panel, and not a product demo. It is a room of people who share a problem, a craft, or a question — and want to think through it together. more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: Birds of Feather (BOF) session
|
Anand S When Data is for Agents, Not HumansWho will be consuming your data - humans or agents? When it’s agents, do you structure it differently? Optimize for token budgets instead of disk or query cost? more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: Hands-on workshop - 2-4 hours
|
|
Anand S Building Verification HarnessesTo confidently deploy in production, you need a robust verification mechanism. Some verification mechanisms are easy. Wrong code doesn’t compile or pass good test cases. Wrong analysis doesn’t meet a post-condition - say a value range or a known aggregate. Wrong proofs don’t validate on LEAN. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Hands-on workshop - 2-4 hours
|
|
AN
Abhijith Neerkaje Workshop instructor AI evals workshopOverview Why do Agents make mistakes - 3 Gulfs [Comprehension, Specification and Generalization]. (10 min) more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Hands-on workshop - 2-4 hours
|
|
UB
Utsab Banerjee Breaking Language Barriers, Not the Bank: Scaling PhonePe’s Aslan to 1.2M Daily QueriesSession Description Traditional keyword search inevitably stumbles when faced with multi-lingual nuances and complex user intent. To solve this at scale, we built and launched Aslan, PhonePe’s natural language search assistant that now seamlessly handles over 1.2 million queries every single day. By breaking language barriers across English, Hindi, Hinglish, Telugu, Bengali, etc Aslan utilizes in… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
|
PC
Pushpendra Singh Chauhan Table-First at InMobi - Migrating a multi-BU data estate to Iceberg "without a big bang"Abstract Most data platforms don’t fail because the tables are wrong — they fail because nobody can find the table, nobody knows who owns it, and every consumer is hard-wired to a physical GCS/blob path that breaks the moment a bucket or partition layout changes. At InMobi, our data landscape grew organically into exactly this: file-and-path datasets with near-zero discoverability, access tightly… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
AB
Aravind Baskaran Self-sufficiency with AIGetting our self-sufficiency up and healthy Self-sufficient systems (and people) are the foundation for us (and any scaleable organisation). Needless to say self-sufficiency is existential for us not just to survive but thrive. Mid-2024 (early AI-wave), we started investing a large share of our time into building the muscle that gets us AI-ready if not AI-native, without hurting our BAU growth. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
u
udit Building Reliable Harness for AI AgentsIn an enterprise the hard part is not making the model clever. Modern models already read code and write decent patches. The trouble starts when we hand an agent real work that runs over many steps and many sessions. It drifts, forgets, and often reports that it is done when it is not. Better prompts do not fix this. What fixes it is the environment we build around the model, and that environment… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
SK
Sanjay Kuniya Defeating the Code Flood: How PhonePe Scales Multi-Agent Loops to Unclog the Dev CycleSession Description As frontier LLMs accelerate code generation across software engineering teams, organizations face a classic realization of Amdahl’s Law. While the code writing phase has been hyper-accelerated, the bottleneck has simply shifted downstream to the sequential, human-in-the-loop code review process. Simply deploying standard AI code review tools or basic chatbot wrappers often exa… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
OS
Omkar Sabade From Figma Node to Production Code: How we Taught Agents to build UI in PhonePeEvery design system has the same expensive seam: the handoff from Figma to code. A designer ships a pixel-perfect screen, and an engineer spends hours inspecting layout, colors, spacing, and variants, then hand-writing it into the codebase over and over, across every platform the design system targets. At PhonePe we built FigGen, an agentic pipeline that turns a selected Figma node into productio… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
|
HS
Hiti Sinha Why Real-Time AI Systems Still Fail: Lessons from Building Decision Pipelines Beyond DashboardsMost enterprise systems today are technically real-time such as, streaming pipelines, event-driven architectures, and low-latency dashboards that are widely adopted. Yet, these systems consistently fail at their primary goal of enabling timely, actionable decisions. In production systems like supply chain and order management platforms, events such as inventory shortages are detected in near real… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
RD
Riya Dhar Too Many Cooks: Production Lessons from Orchestrating Multi-Agent LLM SystemsThe hardest failures in our multi-agent platform didn’t come from the agents - they came from the machinery we built to keep our agents honest. This is a field report from building and operating that orchestration layer in production, where one user-turn fans out across multiple independently deployed agents, gets planned, re-planned, and critiqued before the user sees a token. We’ll set the real… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
RC
Rajmohan C Diagnosing data pipeline failures with AI agents: from research to production, and the open challengesModern enterprise data platforms rely on complex data pipelines for data transformation and integration, with thousands running every day to move data across systems. When a data pipeline run fails, the error you see is usually not the root cause. Failures surface far from where they are born, and the component that throws the error is rarely the one at fault; so diagnosing them means reasoning a… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
MA
Mohamed Ansar Structure Beats Architecture: Lessons from Hierarchical Query Classification for E-Commerce SearchSession Description Search quality in e-commerce hinges on correctly understanding user intent. A query like “cute floral summer dress” looks simple, but mapping it accurately to the right category within a taxonomy of nearly 300 subcategories is a genuinely hard classification problem. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
S
Shivam Your RAG might be three generations behind: a field map of where retrieval is goingAlmost everyone’s RAG journey started the same way: chunk the documents, embed them, retrieve the top-k, stuff the prompt, generate. Then reality hits — semantic drift returns wrong-but-similar chunks, multi-hop questions fall apart because each chunk was embedded in isolation, “summarize across everything” has no good answer, and every attempt to keep data fresh inflates latency. The field has q… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
HA
Himanshu Aggarwal The Graph That Knows What to Watch Next: Real-Time Recommendations with a Content Knowledge GraphCollaborative filtering breaks for new content — there’s no interaction history to learn from. Embedding-based systems improve on that, but they treat items as isolated vectors, blind to the rich semantic relationships between them: shared topics, overlapping entities, genre proximity, mood adjacency. A content knowledge graph makes those relationships first-class citizens and changes what recomm… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
SK
Shril Kumar Cut the Thread: Deleting Customers Without Touching the DataRedacted more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 15 mins talk
|
|
TB
Tanvi Bhakta Cage the coding agent: structure business logic as a harness to subsume LLM-generated complexityPicture this: you’re building a complex pipeline of business logic. Of course, it is 2026, so you’re not writing code by hand. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
Anuj Gupta Building AI on Broken Data: A DataOps Playbook from Processing Millions of Corrupted Data PointsBuilding AI on Broken Data: A DataOps Playbook from Processing Millions of Corrupted Data Points more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
HA
Himanshu Aggarwal From Script to Screen at Scale: Engineering an AI Short Video Generation PipelineGenerating thousands of polished short clips from long-form video — automatically, across multiple content genres — is a different problem from what most AI video demos show you. This talk walks through a production pipeline that does exactly that: automated clipping with LLM-based segment selection, an Intelligent Reframing Engine that detects live speakers vs. static faces using mouth movement,… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
KANIKA SINGHAL When AI Agents Access Your Data: Securing Runtime Flow in Multi-Agent Pipelines{Describe your session in 2 paragraphs} We’re building AI agent architectures that look a lot like distributed systems, passing data across multi-agent pipelines, MCP tools, and external APIs. But because these workflows run autonomously and at incredible speeds, they introduce a brand-new challenge to our standard network models. Once an agent is given tool access to do its job, it executes sub-… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
|
Anuj Gupta AI Makes You Code Faster. Does It Make You a Worse Developer?AI coding assistants are everywhere. Organizations are rolling them out at scale, developers are using them daily, and productivity gains are widely celebrated. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Opening the Virtual Networking Black Box: Building Infrastructure-Aware AgentsAbstract Modern AI agents can process logs, metrics, dashboards, and incident tickets, but often struggle with real-world operational reasoning due to limited understanding of underlying infrastructure. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
Network That Explains ItselfModern enterprise networks generate enormous volumes of telemetry, logs, flow records, and packet captures — yet understanding how traffic actually moves through the infrastructure remains surprisingly hard. Network behavior is hidden behind layers of virtualization, overlays, and cloud abstractions, leaving engineers to manually stitch together signals from multiple tools just to answer basic qu… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
|
SJ
Satyajeet Jadhav  Building an OCR and Data Extraction Pipeline for Business Workflows with LLMsWhile working with our customers in the MSME sector, we realized that most of them rely on WhatsApp for one important reason. It is really easy to capture and send images - invoices, receipts, product photos, attendance, screenshots, etc. A lot of useful business information is trapped inside these images. The problem is that each image is unique and could contain different kinds of information. … more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
A
Apurva From WALs to Indexes: The Database Internals Hidden Inside Modern LakehousesAbstract Lakehouses built on open table formats have emerged as the de facto architecture for modern analytical data systems, yet few practitioners appreciate how deeply database internals underpin their design. Modern open table formats are often described as metadata layers on top of Parquet files, but beneath the surface they have quietly reinvented many of the core ideas that powered database… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
VV
Venkata sai Varada Skills, Native Tools, and MCP: The Architecture Behind an Enterprise AI Agent That Degrades Gracefully and Scales for FreeSession Description Every team building AI agents hits the same wall: the demo works beautifully, then Customer A connects Datadog, Customer B uses Splunk, Customer C has New Relic and a homegrown wiki — and your single agent codebase has to work across all of them without hardcoding anything. This talk walks through the architecture we built to solve this: five layers from user interface to plat… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
NC
Niraj Chauhan Building an AI Copilot for Travel ConsultantsBuilding an AI Copilot for Sales Consultants From phone calls to first-draft itineraries more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
MJ
Mayur Jadhav An Agent That Builds Agents: AI-Powered Recipe Generation for Local-First ETLAn Agent That Builds Agents: AI-Powered Recipe Generation for Local-First ETL more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
h
heetgala Stop Shipping AI on Vibes: AI Evals for AgentsAbout this session Every team building with LLMs has lived the same story. The demo works beautifully, everyone’s impressed, it ships. Then a real user asks a slightly different question and the answer comes back confidently wrong. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Pravein Govindan Kannan Optimized AI Inference with llm-d and a case-study on Sovereign AI on heterogenous GPU clusterIntroducing llm-d llm-d (llm-d.ai) is a high-performance distributed inference serving stack optimized for production deployments on Kubernetes It provides a transparent routing layer that sits between the client and vLLM serving pods, making key scheduling decisions at request granularity: more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
JD
Jaidev Deshpande "Boss, GPU top-up karwa do" - Monitoring Training Costs at ScaleEvery time we ran out of GPU credits, the fix was the same: someone pinged a cloud administrator and said, “Boss, GPU top-up karwa do.” Nobody asked who burned the last batch, on what, or whether the run even finished. That one sentence (which is now a distant memory in our Slack archives) is what a GPU bill sounds like when no one owns the cost. more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
SK
Sujit Kamthe Architecting AI-Ready Enterprise DataTitle Architecting AI-Ready Enterprise Data About In recent enterprise settings, the effectiveness of AI systems is increasingly constrained by data readiness rather than model capability. Although modern enterprises possess substantial data assets, they are often not organized, governed, or operationalized for reliable AI consumption. This session introduces a comprehensive layered framework tha… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
Manu Manjunath Agentic debugging with auto-heal in long-running workflowsDescribe your session in 2 paragraphs This talk shares practical lessons from building an agentic AI system that does deep technical investigations of production failures spanning several services, long-running jobs and data streaming pipelines. Post failure identification, how to apply ‘data fixes’ to mitigate failures, before a permanent fix is rolled out? more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
AS
Amaan Shaikh Your Database Is Downstream of Your ProductSession description For years, our analytics platform served a workflow nobody complained about. We build planning software for out-of-home advertising, the billboards and digital screens you pass on roads, in transit, in malls. Our users are media planners: they assemble a campaign, kick off an insights run, wait, and review. Behind that sat Postgres for analytics, MongoDB and Redis for proximit… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
HP
Hrithik Piyush Many Narrow Agents, Not One Big Prompt: Putting AI on the On-Call RotationAbstract The obvious way to put an AI agent on the on-call rotation is one large prompt that knows everything. We rejected it for a multi-agent architecture: around 18 specialized agents over a large cloud database, a router that classifies each incident to the right specialist, and a shared library of nearly 300 reusable skills. Diagnosis is automated but mitigation stays human-approved. more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 15 mins talk
|
|
FG
FELIX GEORGE Unsupervised Cycle Detection in Agentic ApplicationThe problem being addressed Like traditional software, AI agents are prone to failure — and one of their most insidious failure modes is the repetitive futile cycle: a loop of unproductive behavior in which the agent keeps invoking tools or sub-agents without making any real progress toward its goal. These cycles arise naturally from the plan–act–observe paradigm, non-deterministic tools, error r… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
AM
Arjun Mahishi Bringing Down MTTR: Building an AI-Powered Diagnostic Platform for Database SupportTitle: Bringing Down MTTR: Building an AI-Powered Diagnostic Platform for Database Support Author: Arjun Mahishi (arjun.mahishi@gmail.com; Cockroach Labs) Session type: Talk (30 mins) Track: Building & implementing AI tools & agents in production Submission for: The Fifth Elephant Statue of this doc: Draft (still iterating over it; Will be done before 30th June) more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Vrunda Gadesha A Sovereign Stack for Turning Unstructured Text into Multi Turn ConversationsTraining and aligning enterprise models for regional languages like Telugu is severely limited by text scarcity and the high cost of manual data creation. Without an automated pipeline, creating high-quality conversational data requires hiring bilingual domain experts to manually read documents, extract topics, and draft realistic dialogue trees—a process that is slow, expensive, and difficult to… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 15 mins talk
|
Khushi Deterministic Data Isolation for a Non-Deterministic AgentEvery team has now anchored an LLM onto a database. When we create a conversational agent, in the demo it answers questions beautifully. Then it reaches production, a user rephrases a request, and the agent happily writes SQL that crosses a tenant boundary or reads a table it was never meant to see. This risks data leaks, privacy and security threats. The first few lines of treatment would be to … more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
PP
Purushotham Pururava Pushpavanth The Missing Half of AI Data Assistants: A REPL for PipelinesHow Nexus-AI closes the loop on AI-generated pipelines — running the real job and proving the output is correct. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Swetha A 
Mahita D  From Amnesia to Photographic memory in Agentic SystemsObjective / Abstract : You’ve built a few agents, wired them together into a multi-agent system. What comes next? This workshop takes a focused look at one of the most critical and often underestimated parts of an agentic system: memory. Participants will come away with a practical understanding of how memory shapes agent behavior, capabilities, and user experience. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Hands-on workshop - 2-4 hours
|
|
NA
Navdeep Agarwal From PDF to SQLFifth Elephant 2026 Submission Description Most “ask my documents” systems stop at retrieval: upload PDFs, create embeddings, and search for relevant chunks. That works when the user wants surrounding context, but it starts breaking down when the user needs numbers for a report, dashboard, analysis, or presentation. In Indian mutual fund reports, the useful answers are usually structured facts: s… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
Anay Nayak Shipping an MLOps Platform: What we let the AI own and what we didn'tWe built a production ML forecasting platform under a hard deadline with significant instability underneath it: more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
VN
Vikram Nayak When There's No Unit Test for "Good": A Maker-Checker Loop for Subjective AI OutputBRIEF DESCRIPTION: The problem: When an AI agent writes code, you can test whether the code works. But when an agent makes a chart, how do you test whether it’s any good? A chart can be technically correct and still fail to get its point across. “Good” depends on the audience and the decision they need to make - there’s no test that returns true or false. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
AN
Aayush Naik Delta Lake Write Internals: INSERT, UPDATE, DELETE From the Ground UpDelta Lake makes the table mutable, but the underlying parquet files are physically immutable. In this talk, we will dive into the internals of Insert, Update and Delete operations. We begin with the introduction: Parquet (columnar storage), Delta log (txn record of all add & remove actions), and define that these constitute the Delta Table. We begin with INSERT, which is straightforward, write a… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
PJ
Prasanth Jayaprakash From Scratch to Production in 3 Months: The AI-First Delivery ModelMost teams bolt AI onto existing processes: same Jira workflows, same sprint rituals, same onboarding docs, just with Copilot autocompleting some code. But what if you could start fresh? What if you had a greenfield project and could design your entire delivery process knowing what AI can do today? more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Zahle Making Generative UI Work in ProductionDescribe your session in 2 paragraphs Generative UI changes what an agent can return. Instead of answering with a wall of text, the agent can render charts, forms, workflows, tables, dashboards, and other interactive interfaces. That looks great in demos, but production exposed a harder question for us: what should the model actually emit to make this reliable, fast, and streamable? more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
AR
Anusha Rao Beyond Text-to-SQL: What "Agentic-First" Really Means for a DatabaseBuilding a Database for the Agentic Age Session title: Beyond Text-to-SQL: What “Agentic-First” Really Means for a Database more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 15 mins talk
|
Hritik Raj From Stateful to Stateless: Evolution of the Model Context Protocol (MCP)The “why”? Every decade or so, a protocol comes in and changes everything. The Model Context Protocol (MCP) was one such thing which took the place of this decade and which became the momentum to AI. Since its first specification release, it has fundamentally reshaped how AI models talk to the world: external tools, live data, real services. But here’s what nobody tells you about the quiet revolu… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
VK
Vinayak Kadam Structured Extraction in Production: State, Schemas, and the Engines Underneath.Abstract Structured extraction from unstructured text is often presented as a solved problem, enabled by LLM frameworks, schema libraries, and built-in tooling. In practice, however, production systems expose a different set of challenges around scale, multi-turn interactions, reliability, and long-term maintainability. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
GM
Gaurav Maheshwari From ReAct to Multi-Agent System - Our Journey of building an Observability agentAt Oodle AI, we have been building AI agents for faster incident debugging, helping engineers with day-to-day oncall/operational tasks. Our journey started about 1.5 years back with a Langchain ReAct based agent. Today, we are running multi-agent orchestration to allow our customers to just ask questions on top of their observability data. This session will go over our learnings in running this s… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Karthika Vijayan When the Orchestrator Wasn’t Enough: A Production Story of Building Reliable Multi-Agent AIWe built a production multi-agent assistant for a complex enterprise planning workflow where AI assists, rather than replaces, human decision-making. Users express their intent through natural language, the system extracts structured semantic information for human review, and multiple specialist agents collaborate to retrieve data, perform analysis, and generate recommendations. Our initial archi… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
shubhankar khare Harness Engineering: Build a Minimal Coding Agent from ScratchAbstract Everyone can get an AI agent working in a demo. Keeping it working - and understanding why it behaves the way it does, is a different skill. And it turns out that skill has surprisingly little to do with the model. It’s about the harness: everything you build around the model to turn it from something that talks into a system that acts. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Hands-on workshop - 2-4 hours
|
|
SK
Sujit Kamthe Agents Don’t Fail Because They’re StupidAt some point, many agentic systems end up with a prompt that nobody wants to touch. What began as a few instructions gradually accumulates incident fixes, business rules, workflow logic, examples, exceptions, and model-specific caveats. Over time, the prompt becomes one of the most critical pieces of software in the system despite rarely being designed, reviewed, or maintained like software. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
PM
Pedro Mázala Presenter Your Agent Is Just a Stateful Consumer (Don't Tell It)Technologies: Apache Fluss, Apache Flink, Apache Iceberg more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
MS
Mayur Singal Beyond Metadata: Building an Open Context Layer for AIDescribe your session in 2 paragraphs Everyone has seen how AI transformed software engineering. Tools like Claude Code and Codex work because code comes with rich context—source control, dependency graphs, reviews, tests, and execution history. Enterprise data, however, lacks an equivalent foundation. AI agents are expected to answer questions, generate SQL, understand business metrics, and auto… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
PB
Pranav Balasaheb Bhosale Vector DBs Are Overrated: Grounding LLM Agents on Master Data Without the OverheadLarge Language Models are powerful at reasoning and recommendation, but they routinely hallucinate entities that do not exist in an organization’s proprietary data. The common response is to build a Retrieval-Augmented Generation (RAG) stack with embeddings, a vector database, and retrieval infrastructure. However, many enterprise datasets are not collections of long documents—they are structured… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
|
SK
Santosh Kewat Your AI Agent Is Guessing. Give It Context It Can Trust.Session Description Everyone is plugging LLMs into their data right now. You wire up an MCP server, point Claude or Cursor at your warehouse, ask “what was revenue last quarter?” and get back a confident, well-formatted, wrong answer. The model picked the staging table. It used a deprecated column. It invented a join. It had no idea your team redefined “active user” six months ago. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
Khushi End-to-End Observability for AI SystemsSpeaker bio I am Khushi, working as Lead ML Architect in a fintech-space company called Finantic.AI. I am enthusiastic about new technologies built on top of existing fundamental technologies that work just a little better. more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: Hands-on workshop - 2-4 hours
|
|
KB
Kusumakar Bodha From Documents to Data: Tiered Extraction at Enterprise ScaleDescription At a handful of documents, extraction is trivial — you upload them to ChatGPT and start asking questions. Across a production corpus of millions of documents — images, PDFs, spreadsheets, slide decks, office files — that grows in bursts as each new customer onboards, it becomes a data-engineering problem: turning every kind of messy enterprise file into clean, retrievable data, reliab… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 15 mins talk
|
Akash Sathish Before the Agent Calls exec(): Source-Level Findings from 100 MCP ServersMost teams adopting MCP servers treat security the way early npm treated dependencies, install, trust, ship. When I built MCPeek, an AST-level static analysis tool, and pointed it at 100+ popular open-source MCP servers, the results were uncomfortable: 445 real findings across 70 servers which carried at least one exploitable pattern: command injection from tool input, path traversal, SSRF, and t… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
Akash Sathish Preserving Human Intuition in AI DevelopmentThree major studies from Anthropic, METR, and MIT recently revealed an uncomfortable reality about AI-assisted development. Their data shows that AI tools can lower skill assessment scores, increase actual task completion times, and significantly reduce neural connectivity during creative problem-solving. This is not a simple case of developers becoming lazy. Instead, widespread AI usage actively… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Birds of Feather (BOF) session
|
|
SB
Sourav Bhuwalka Building Realtime CDC and Fabric Mirroring(Streaming data) at scale : Solving Replication Lag and Schema Evolution in Real-Time Data Platforms{Describe your session in 2 paragraphs Real-time analytics platforms promise fresh data without complex ETL pipelines, but operating them at cloud scale introduces a very different set of challenges. In this talk, we share lessons learned while building Microsoft Fabric Mirroring, a system that continuously replicates Azure SQL Database workloads into OneLake with near real-time latency eliminati… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
AR
Aishwarya Raimule Leverage Envoy AI Gateway To Operate MCP Servers At ScaleAs organizations adopt AI agents, the number of Model Context Protocol (MCP) servers inside the enterprise is growing rapidly. Teams are deploying MCP servers for GitHub, Jira, Kubernetes, observability platforms, internal APIs and business systems. While MCP standardizes tool integration, operating dozens of MCP servers introduces new challenges around authentication, authorization, discoverabil… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
Bibhas Debnath Decentralized Ownership, Centralized Control: Building a Self-Healing Data PlatformBrief abstract Bringing together all the org data by following a data mesh architecture in a decentralized manner sounds great at start, but the approach quickly falls on its face because of strict data governance policies around cross-team data access. Getting governance team approve access request for hundreds of teams get real painful and slow immediately, and the data product teams start fall… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|
|
NS
Narayana Sastry Submitter Test Every Idea, Together: An Autonomous ML Agent That Multiplies Team Research BandwidthSession Description Research teams are bottlenecked on the number of experiments they can run. A team’s research output can be roughly formalized as (# of people) x (# of experiments) x (research taste). With fixed headcount, the question becomes: how do we reduce the marginal cost of running one more experiment, without lowering the quality of judgment behind what gets tested? more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
S
Sathish Presenter AIOps: Leveraging AI for Software IncidentsDescription During production incidents or on-call schedules with a barrage of alerts, engineers must sift through hundreds or thousands of services, code changes, metrics data points, logs, and traces to reason about the issue and find the root cause. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
SR
Shashank Rao Facilitator Beyond SWE-bench: How do we evaluate AI agents for real-world workflows?Track: Track 2 – Building & Implementing AI Tools & Agents in Production Format: Birds of a Feather (BoF) Session more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Birds of Feather (BOF) session
|
Prateek Mandloi How Should We Engineer Agent Quality Loops?Description Many teams now have traces, logs, evaluations, user feedback, and dashboards for their AI agents. The harder question is what those signals should prove, and what teams should change after seeing them. more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: Birds of Feather (BOF) session
|
|
Prateek Mandloi Submitter Agents Leave TracesDescription Agents are often discussed as if the model is the whole system. In practice, production agents behave more like running programs: they observe, decide, call tools, read outputs, recover from errors, and continue. If that loop is the program, the trace is its stack trace. This talk is a concrete engineering story about using turn-level traces to improve agent quality. I will show how w… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
SR
Shashank Rao Submitter Building an Agentic Request Resolution SystemDescription Unlike coding or conversational question answering, support request resolution has no generic workflow that an AI agent can follow. Resolution paths vary across customers and domains (for example, using Okta vs. IdentityNow for access requests), involve constantly changing context, require high-risk actions (such as granting admin access or removing users from licenses), and must stri… more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 30 mins talk
|
|
SJ
Sparsh Jain  Two Years of Experience, Zero Learned: Here's How to Fix ThatDescribe your session in 2 paragraphs I have worked with my Al agent for two years. I gained two years of experience. My agent gained zero. That same agent worked with ten people on my team. Combined how can we give our agent 10-20 years of experience. How do we design this? more
I am submitting for: Track 2 - Building & implementing AI tools & agents in production
Type of session: 15 mins talk
|
|
GM
Garima Mishra Full Refresh to Incremental: Rebuilding Denormalization for Reporting at Razorpay ScaleAbstract / Session Description At Razorpay, our lakehouse platform ingests over 6 billion events daily and powers a reporting platform that generates close to a million reports every month. As scale grew, full-refresh denormalization became unsustainable: joins across 10-30 entities consumed heavy compute, report freshness lagged by up to 48 hours, and highly mutating datasets made derived tables… more
I am submitting for: Track 1 - Data engineering & infrastructure
Type of session: 30 mins talk
|