Research Scientist & Principal Architect
Cognizant AI Lab
San Francisco, CA · Dec 2018 – Present
Lead Researcher for Multi-Agent Architectures and Evolutionary Optimization.
Caesar: Autonomous Reasoning & Knowledge Synthesis
- Architected Caesar, an autonomous research agent comparable to Gemini's Deep Research. Scaled inference-time compute via an adversarial refinement loop that critiques internal drafts, generates orthogonal queries to attack weaknesses, and consolidates findings through generative merge.
- Replaced static RAG with a dynamic Perceive-Think-Act loop backed by a self-organizing knowledge graph, enabling associative reasoning that surfaces non-obvious cross-disciplinary connections.
- Designed an active information-foraging policy that detects stagnation and autonomously executes strategic backtracking to ensure open-ended learning over long horizons.
Multi-Agent Systems & Large-Scale Evolutionary Optimization
- Developed an FSM-based multi-agent code-generation framework, evolving the composition of specialized LLM teams to achieve strong performance on SciCode and HumanEval+.
- Designed a hierarchical expert-agent pipeline that decomposes hard coding problems and routes sub-tasks to specialized LLM roles.
- Pioneered Evolutionary Population-Based Training (EPBT), evolving hyperparameters and loss functions during training to reduce compute relative to grid search.
- Pioneered evolutionary prompt optimization, improving LLM performance on challenging code-generation benchmarks.
- Created Code Archaeologist, an agentic system that reasons over Git/LFS repository histories to detect architectural patterns and technical debt.
- Scaled distributed ML systems to produce SOTA neural networks for vision and language tasks.
Research Scientist
Sentient Technologies
San Francisco, CA · Jun 2017 – Nov 2018
- Invented and scaled CoDeepNEAT (evolutionary neural architecture search) to clusters of hundreds of GPUs, demonstrating that evolutionary search scales with parallel compute.
- Achieved SOTA results on image captioning and object recognition by evolving topologies that share weights across tasks.
- Designed and deployed production AutoML systems for real-world use cases.
Research Intern
Sentient Technologies
San Francisco, CA · Dec 2015 – May 2017
- Researched early applications of evolutionary algorithms to deep neural networks, laying the groundwork for CoDeepNEAT.
- Built scalable ML infrastructure for distributed neural-architecture optimization.
Research Assistant / Ph.D. Candidate
The University of Texas at Austin
Austin, TX · Aug 2013 – Dec 2018
- Developed MEA (Meta-Evolutionary Algorithm), a bilevel optimization framework, applied to non-differentiable continuous-control tasks such as helicopter hovering where standard gradient-based RL struggles.
- Created the original CoDeepNEAT algorithm, combining evolutionary computation with hierarchical coevolution to discover deep neural architectures.
- Published extensively in top venues including GECCO and Applied Soft Computing.
Research Intern
Open Source Robotics Foundation
Mountain View, CA · May 2015 – Aug 2015
- Contributed to the Gazebo 3D robotics simulator and authored a RoboCup plugin for 3D soccer simulation.
Research Assistant
UC Berkeley
Berkeley, CA · May 2012 – Aug 2013
- Built a computer-vision system for vision-based indoor localization from cellphone imagery, fused with LIDAR-derived point clouds.
Intern
Qualcomm
Berkeley, CA · Sep 2011 – Dec 2011
- Built benchmarking and performance-measurement tools for AR applications (Layar SDK) on Android.