Flatiron Institute / Polymathic AI
Assistant Research Scientist · New York, NY Sep 2026 – Present
- Benchmarking and optimizing inference pipelines for the 1.28B-parameter Polymathic Walrus multimodal scientific foundation model.
- Built automated optimization pipelines that cut peak GPU memory by 47% (6.18 GB → 3.27 GB) and accelerated sample throughput by 2.3×.
- Maintained an 82.3% mean accuracy (within 0.2% of the FP32 baseline) through 50% model-footprint compression.
Siemens
Multi-Agent Software Architectures Researcher · Princeton, NJ Jun 2026 – Sep 2026
- Architected a distributed recursive task decomposition algorithm for multi-agent reasoning, decomposing complex intents into Directed Acyclic Graphs (DAGs) with per-path blackboards across heterogeneous LLMs.
- Achieved up to 36× faster intent resolution and 25× token reduction versus centralized orchestrators (Qwen, Claude Opus-5, Claude Sonnet-4.6).
- Maintained 100% task correctness and completeness across 22 synthetic intents on a testbed of 4 MCP-conformant microservices.
University of Iowa
Graduate Teaching Assistant — Software Engineering & Java · Iowa City, IA Jan 2026 – May 2026
- Instructed and mentored 100+ students in object-oriented design, MVC/Singleton patterns, multi-threading, and backend Java development.
- Led code reviews and lab sessions on TDD, JUnit, and automated CI/CD grading pipelines under Agile sprint cycles.
National Renewable Energy Laboratory (NREL)
Research Scientist — Data & Compute Systems · Golden, CO May 2021 – Oct 2025
- Implemented Graph Neural Networks for structural data compression, improving simulation throughput by 50% with no accuracy loss.
- Architected a modular deep learning benchmarking framework processing a 1.7 TB dataset across 483,000 hyperparameter configurations and 11M training runs.
- Engineered compression pipelines (Iterative Magnitude Pruning, activation variance) compressing networks by up to 90% while preserving accuracy.
- Refactored legacy Fortran90 codebases for scalable HPC deployment (OpenMP, MPI), cutting runtime by 30%.
Cornell University
Research Scientist — AI Systems & Computational Biology · Ithaca, NY Jun 2022 – Oct 2022
- Built scalable deep learning pipelines over epigenomic datasets for sequence-free enhancer prediction and chromatin architecture identification.
- Published APSO (Adversarial Particle Swarm Optimization) in Bioinformatics (2025), an architecture-independent XAI and adversarial robustness framework.
Los Alamos National Laboratory (LANL)
Post-Baccalaureate Researcher / Research Software Engineer · Los Alamos, NM Jul 2018 – Aug 2020
- Applied heuristic optimization (Particle Swarm, Gravitational Search) to calibrate nonlinear multi-physics simulations against empirical data.
- Modernized CI/CD and automated regression testing frameworks for massive-scale multi-physics models.