Abstract

AI is moving beyond code completion to act as a true co-developer across the software development lifecycle (SDLC). From structuring raw data into validated schemas to guiding analytical reasoning with statistical evidence, AI can reduce friction in everyday developer and data science workflows.

This talk showcases two open-source experiments that demonstrate this shift:

• SchemaForge: Goes beyond schema inference by creating ready-to-use DBT models, test rules, and ER diagrams. It also supports Python-based ETL pipelines, making schema outputs directly usable in production workflows.

• Hypothesis Forge: Helps data scientists move from questions to evidence faster. It analyzes datasets, suggests testable hypotheses, runs statistical checks, and summarizes results into clear, interpretable insights.

Together, these tools illustrate a unified vision of AI-augmented development—from structuring, validating, and ingesting data (SchemaForge) to generating and testing hypotheses with statistical rigor (Hypothesis Forge). Through live demos, we’ll explore how AI can ground generative outputs in structure, validation, and insight—and what this means for the future of AI-assisted workflows.

What the audience will take away

• How AI can generate DBT-ready rules from raw data, with optional Python ETL ingestion support.
• How automated hypothesis generation + statistical testing can guide faster, evidence-based analysis.
• Patterns for integrating generative AI into real-world SDLC stages beyond “autocomplete.”
• Lessons from building open-source AI tools: trade-offs, reliability concerns, and guardrails.

Format

• Duration: 30 minutes
• Type: Experiential talk with live demos
• Structure:
  1. Framing: AI in SDLC beyond code completion (5 min)
  2. Demo: SchemaForge — schema inference → DBT rules → ingestion → execution (10 min)
  3. Demo: Hypothesis Forge — hypothesis generation, testing, synthesis (10 min)
  4. Reflections and Q&A (5 min)

Target Audience

• Data engineers and analytics developers who deal with repetitive schema, ingestion, and pipeline tasks.
• Data scientists who want to test ideas quickly and systematically.
• Developers interested in building or contributing to AI-assisted open-source tools.

Speaker Bio

I’m Prudhvi Krovvidi, a Data Scientist at Gramener, where I explore how AI can simplify and accelerate data workflows. Most of my work ends up as open-source experiments on GitHub — from schema inference and quality checks to decision tree generation and AI-assisted analytics. I enjoy building lightweight tools that bring AI into everyday developer and data science tasks — and when I’m not doing that, you’ll probably find me out on my bike.

Company: Gramener, Hyderabad
GitHub: github.com/prudhvi1709
Email: kprudhvi71@gmail.com
Linkedin : Prudhvi Krovvidi

Links

• SchemaForge
• Hypothesis Forge