AI Systems Architect & Research Engineer
Founder, Syntropy Systems · Nashua, NH · Open to AI Systems Architecture, MLOps, AI Safety, and Product roles
10 years building data and ML systems — from high school SQL dashboards to production infrastructure handling 525K requests/month. Research background: 70-paper literature synthesis on AI safety, backpropagation implementation from 1986 Rumelhart paper, research mentorship designing paper curricula. Architected sub-1s semantic search across 300k documents using vector embeddings + MongoDB hybrid queries. Created Agent Tokens Protocol (open standard for AI agent identity). Built 5-agent orchestration pipeline handling 150k+ line edits with zero regressions. WPI Data Science graduate (3.75 GPA, High Distinction). BlueDot AI Safety Alignment certified.
Research paper exploring a gap: do models actually believe what their creators say about safety? I call this "meta-alignment"—not whether models do safe things, but whether they've internalized their lab's safety philosophy. Surveyed 10 frontier models with 40 questions derived from 70+ papers. Interactive tool included.
A framework that makes data science experiments reproducible. Jupyter notebooks hide state—you can't tell what order cells ran or what values held. Crystallize treats each experiment as an immutable record with automatic statistical checks. Long-term goal: infrastructure that lets AI agents run their own experiments.
A protocol for declaring agent intent at the HTTP layer. The problem: when an agent makes a request, origins can't tell if it matches what the user actually wanted. User says "check weather," agent calls the bank API—how does the bank know to block it? Agent Tokens let agents declare their allowed scope upfront so middleware can enforce policy automatically.
A proposed composition layer for MCP that lets agents submit multi-step plans as a single typed, auditable expression — without routing intermediate data through the model. Built entirely from already-specified MCP primitives, so it runs on top of the existing spec rather than forking it.
Minimal replication of Anthropic's Golden Gate Claude on consumer hardware. Trains a Sparse Autoencoder on Qwen2.5-1.5B, discovers interpretable features, and steers model behavior — all on an RTX 3070 Ti. Demonstrates that mechanistic interpretability research is accessible beyond frontier-scale compute.
Fork of the companion code for "Verbalizable Representations Form a Global Workspace in Language Models" (Transformer Circuits), adding a backward "ghost-token" lens — the original Jacobian lens run in reverse, decoding residual states back into input-token space.
Adversarial token convergence experiments on MLX. Pits language models against each other to reveal training strata — solo mode converges in 2 iterations, adversarial mode never does. Under pressure, models collapse to their most defensible tokens, revealing composition invisible in normal evaluation.
Unittest for tool-using LLM agents. You don't fly a new airframe straight into a storm — Wind Tunnel is a reliability bench where agents run reproducible scenarios scored on four independent layers: outcome, trajectory, constraints, and robustness. Structured, diff-able, and runnable in CI. Open source from Syntropy Systems.
Event-driven ML service architecture enabling long-running agentic workflows. Designed around distributed systems constraints: Kafka for async (no HTTP timeouts), Redis for large results (Pusher limits), dual-mode for internal app + external API. 525K requests/month in production.
Addressing the "Translation Gap" between users and engineers. An intelligent pipeline that converts vague bug reports into structured, high-entropy technical tickets using LLMs.
Your identity as an API endpoint. An MCP server that lets AI agents query your professional profile with structured tools instead of scraping HTML.
Local experiment orchestration: queue jobs, track metrics, wake up to results. A lightweight, self-hosted alternative to cloud experiment trackers — no accounts, no external dependencies, just SQLite and your filesystem. Open source from Syntropy Systems.