Engineer · Researcher · Essayist
I build the agentic evaluations that measure Gemini. I write about the agents that will replace it.
§ argument · 2008 → 2026
“Complex, intelligent behavior emerges from simple rules applied at scale — the same principle behind modern distributed AI systems.”
§ writing
Inside the Evolution Arena and the rise of autoresearch.
How particle swarm optimization escapes the local-minima trap in ARC-AGI.
Local Tinker — a clean API for local LLM fine-tuning.
§ tutorial · interactive encyclopedia
Machine Learning, Visualized. A visual encyclopedia of the ideas behind modern ML.
Editorial playgrounds, one concept at a time — starting with the linear-algebra foundations (matrix transformations, eigenvectors, PCA, SVD) and working up through gradient descent, regression, neural networks, attention, diffusion, and beyond.
§ selected work
Google Cloud · 2025 → present
Defining the tasks, building the datasets and implementing the scoring that measure agentic capability on Gemini checkpoints — integrated into Denigma, Google’s model-evaluation platform, for headroom measurements and head-to-head model comparisons that research leads use to prioritize capability work.
Google Cloud · 2019 → 2025
Led backend storage and data integrity for CrowdCompute, Google’s large-scale human-data engine generating RLHF and SFT datasets for foundation models. Architected DQaaS, a low-latency LLMOps service (gRPC / Protocol Buffers) for enterprise-scale prompt versioning, testing and retrieval in GenAI and RAG workflows.
Amazon · 2017 → 2019
Architected and launched a centralized ingestion platform processing 10B+ events/day, with fault-tolerant streaming pipelines on Kinesis, Lambda, and S3 under on-call ownership. The data lake's front door.
IISc · 2006 → 2008
Master's thesis on multi-source odor localization with swarm robotics — the work behind US Patent 8,838,271 for nuclear spill detection, and the conceptual scaffolding behind the multi-agent reasoning work I'm doing now.
§ the record
Also: Senior Member of IEEE · FBCS · Hackathon Raptors Fellow · IEEEXtreme Proctor · GATE 99th-percentile scholar · Utah State Grad Research Scholarship · 21 internal Google awards, including 5 spot bonuses. Mentor/TA at Google Tech Exchange; mentor at Founder Institute and Opportunity Machine.
Multi-agent systems for the ARC-AGI challenge — abstract reasoning frontier.
Pythonarc-agi-searchSwarm-based search applied to ARC-AGI — bio-inspired optimization for reasoning.
Pythonrecursive-self-improvementRSI framework for LLMs — iterative capability enhancement.
PythondeepsnakeDeep Q-Network agent that learns to play Snake from raw reward.
I started as an aerospace engineer designing intake systems for missiles, became a swarm-robotics researcher under Debasish Ghose at IISc, picked up a US patent and a Cornell MEng on the way, and have spent the last decade shipping data infrastructure at Amazon and Google.
Today I build post-training agentic evaluations for Gemini, after six years on the data engine behind Google's foundation models. Off-hours, I write essays and run experiments on multi-agent reasoning — the same swarm research continued in a new vocabulary.
Saratoga, California · Senior Member, IEEE · FBCS