About

Hi, I am Madhava Gaikwad.

I am a Senior Software Engineer working at the intersection of distributed systems, AI alignment, privacy and differential privacy, and ML systems. I turn theory into runnable, tested artifacts.

I work on bridging AI alignment theory with real-world systems. My background spans differential privacy, reinforcement learning from feedback, distributed consensus, and novel data abstractions for telemetry and streaming analytics. The questions I keep coming back to are these. How do we reason about uncertainty rigorously. How do we guarantee safety and rollback in complex infrastructure. How do we design data and learning pipelines that stay robust at scale.

What I build

Small, composable systems. Clear quickstarts. Minimal tests and CI. Figures you can reproduce.

Focus areas

  • Alignment loops, including RLHF and beyond
  • Operator views of learning
  • Differential privacy for LLMs
  • Reliability at datacenter scale

Talk to me about

Turning research into demos. Privacy-first LLMs. Robust infrastructure for ML.

Contact