Founder Profile

Deepak Karkala, Founding AI Engineer at Vavi Labs.

Senior ML Engineer building AI-native products, dev tools, plugins for coding agents, fine-tuned models, and technical education systems. Vavi Labs is the lab where this work comes together across products, consulting assets, explainers, and applied AI projects.

Current work.

Four workstreams: governance product, interview-prep platform, dev tooling plus explainers, and fine-tuned model projects.
01

Arbiter

AI governance and decision-intelligence platform for engineering teams, currently in beta testing.

  • CLI-first engineering workflow
  • Audit trails and confidence scoring
  • Connector-backed decision context
  • Async review loops for low-confidence decisions
Open Arbiter →
02

Tech Abstractions

AI/ML interview-prep platform focused on production reasoning, system design, LLM infrastructure, and applied practice.

  • Few dozen users on the waitlist
  • Structured learning paths
  • Workbook-style practice
  • Expert walkthroughs and applied exercises
Open interview prep →
03

Dev Tools & Explainers

Dev tools, plugins for coding agents, and visual explainers for agentic AI engineering and production MLOps.

  • Plugins and workflows for coding agents
  • Agentic AI engineering patterns
  • Production MLOps workflows
  • Illustrated guides for LLM infrastructure
Browse dev tools →
04

Fine-tuning Projects

Focused model projects that combine custom datasets, training workflows, evaluation harnesses, and domain-specific generation.

  • Sitcom Scriptwriter for screenplay-style generation
  • Kannada Physics Tutor for local-language learning
  • Small language model fine-tuning
  • RAG and evaluator components
View GitHub →

Background: 6+ years in ML/AI across production e-commerce ML, IoT anomaly detection, energy forecasting, GenAI, RAG, fine-tuning, and MLOps.

Operating style: evaluation-first engineering, cost-aware architecture, practical governance, and direct translation from product problem to shipped AI system.