Who is your stack really built for now?

Built for humans. Now rebuilding for agents. The Fifth Elephant 2026 Annual Conference - from production, not theory.

For years, data platforms, voice interfaces, inference pipelines, and testing frameworks were built for human assumptions — human-readable schemas, human-paced conversations, human-reviewed test cases. Agents are quietly breaking every one of those assumptions.

The Fifth Elephant 2026 brings together practitioners who have already hit these walls in production — at Dream11, IDfy, Onehouse — and rebuilt from first principles, with the numbers to prove it.

Two tracks, one question: what does your stack look like when agents, not humans, are the primary users of it? The Fifth Elephant 2026 brings stories from production and the trenches.

📅 31 July 2026 | 📍 NIMHANS Convention Centre, Bengaluru

⏰ One ticket. Five weeks of workshops + the main conference. Price goes up every Friday until 8 August.

Get your ticket → https://hasgeek.com/fifthelephant/2026/#tickets

Includes all pre-conference workshops (7–25 July, plus 8 August) and the main conference day (31 July).

🔴 COMING UP — 24 JULY

Stop guessing. Start measuring. - with Abhijith Neerkaje

Why do agents fail — and how do you actually measure it? This hands-on workshop covers the full loop: designing metrics that matter, generating synthetic test data at scale, running error analysis, and building evaluation pipelines you can put into production.

Two hours, hands-on. You’ll leave knowing how to systematically evaluate agentic systems — not just eyeball outputs and hope. Familiarity with Python required; bring access to Google Colab.

Workshop details →http://has.gy/rgWP
Get your Fifthel ticket to attend → https://hasgeek.com/fifthelephant/2026/#tickets

Part of the Agentic AI - Platforms and Use Cases track at The Fifth Elephant 2026

Conference tracks

Track 1: Data Engineering & Infrastructure

Data platforms were built for a human at the other end — cleaned, joined, dashboard-ready. That assumption is breaking. Agents read messy data differently, pay by the token instead of the query, and increasingly operate the platform itself rather than just querying it. This track brings practitioners who’ve rebuilt data infrastructure from first principles for this shift — from serverless compute that beats the ease/performance/cost trilemma, to hands-on work on what “agent-ready” data actually looks like.

— Curated by Piyush Goel (Razorpay), Ravi Balgi (Datanimbus), and Ranganadh Thata (Mico)

Track 2: Agentic AI - Platforms and Use Cases

Building for agents in production breaks assumptions engineers didn’t know they were making — that voice interfaces are just chatbots with audio, that GPU cost is a fixed tax, that testing an agent works like testing regular software. This track is about what actually held up once these systems hit real users at scale — with the production numbers, failures, and fixes to back it up.

— Curated by Jagadish Kasi (Tryft) and Yash Gandhi (OrcaSheets)

Featured workshops and talks

Track 1: Data Engineering & Infrastructure

🔴 Workshop — 7 July · When Data Is for Agents
S Anand, Straive
Hands-on: rebuild a real dataset for agent consumption instead of human consumption — then use agents themselves to do it.

Talk · Breaking the Trilemma: Serverless Data Platform in Your Own Cloud Account
Nilesh Mahajan, Onehouse
Ease-of-use, performance, cost — pick two, or engineer your way to all three. A working answer to a problem most data platforms haven’t solved.

Track 2: Agentic AI - Platforms and Use Cases

Talk · Your Voice Agent Is (Probably) Doomed
Dhruv Nigam, Dream11
Five months and 8 discarded prototypes later: why treating voice agents as chatbots-with-audio is the wrong playbook, and what Dream11 built instead for 5M+ daily users.

Talk · Beyond GPUs: Cutting ML Inference Costs by 10× Without Sacrificing Latency
Vivek Kalyanarangan, IDfy
GPU cost isn’t fixed — how a 40-model production fleet cut cost-to-serve 10x at the same latency, and the playbook to repeat it.

🔴 Workshop — 17 July · AI Evals: Stop Guessing, Start Measuring
Abhijith Neerkaje, Beyond Vectors AI
Two hours, hands-on: why agents fail differently from regular software, and how to build evaluation pipelines that actually catch it.

Who should attend, and why

This isn’t a conference for people curious about AI in the abstract. It’s for people already responsible for systems that are starting to break under agent-driven assumptions:

• Data engineers and platform teams rethinking how data is structured, stored, and served — now that agents, not analysts, are often the ones querying it
• ML and AI engineers building or scaling agents, voice interfaces, or inference pipelines in production — and hitting walls the standard playbooks don’t cover
• Engineering leaders and architects making infrastructure and cost decisions that will age badly if they’re built on human-era assumptions
• QA, evals, and product folks figuring out how to test and measure systems that don’t behave like traditional software

If your team’s stack was built for people and is now quietly serving agents too, this is where the people solving that problem in production are talking about it.

What does your ticket include

Your ticket covers the entire Fifth Elephant 2026 season — not just one day:

• 🔴 7 July — When Data Is for Agents (S Anand)
• 🔴 24 July — AI Evals: Stop Guessing, Start Measuring (Abhijith Neerkaje)
• 🔴 25 July — Additional workshops (details announced soon)
• 🔴 31 July — Main conference day, NIMHANS Convention Centre
• 🔴 8 August — Closing workshop block (details announced soon)

Volunteer with us

The Fifth Elephant runs on more than speakers and sponsors — it runs on volunteers. We have 20 roles open across the event, from floor operations to speaker coordination, with tiered deadlines depending on the role.

See open volunteer roles →

Attend

🎟 Conference tickets and annual memberships

Conference ticket

• Cancel any time up to 24 hours before the event for a full refund, no questions asked — this doesn’t apply once you’ve attended any workshop under the ticket (7, 17, or 25 July); attending a workshop locks the ticket.
• Includes access to all Fifth Elephant activities — workshops, meetups, and conferences — for one year.
• Easier to get reimbursed through company/HR.

Annual membership

• Access to all Fifth Elephant activities — workshops, meetups, and conferences — for one year.
• No cancellation.

Sponsorship inquiries

Write to sales@hasgeek.com

Contact

📞 Call or WhatsApp: +91 76763 32020
📧 info@hasgeek.com