Portfolio · Selected work
Eight #1 positions on the HuggingFace Open LLM Leaderboard, competing with major tech firms and AI labs using original post-training techniques.
7+ PRs merged into mainline Kubernetes, Argo, Atlantis, and SurfSense — addressing real production-scale problems.
Compact neural vision codec and visual tokenizer — 16:1 spatial compression at 97.69% fidelity, under 150K trainable parameters, batches 6× 40MP images on a single RTX 4090.
Classification, retrieval, and NLP tasks from LLM embedding geometry — only lightweight forward pass components, 28x faster than conventional transformers.
Cybertron 7B v2 hosted as a first-party model in Cloudflare's Workers AI catalog — the only third-party fine-tune in the lineup, served at the edge for nearly two years.
An auxiliary loss-based architecture patch for HuggingFace Transformers, applied during SFT/RLHF. 18 public releases across multiple base models, with multiple #1 leaderboard positions.
A spec-driven agentic ecosystem for long-horizon engineering on enterprise brown-field code.
Regularization technique using Gumbel-sampled noise during SFT/RLHF. Combines with UNA (UNAMGS) for additive performance gains.
Exploratory parameter-efficient adaptation that competes with LoRA on GLUE at <0.25M trainable params, with zero-overhead expert switching at inference.
Redis-first, event-driven workbench with swarm intelligence for long-running Claude coding sessions. JSONRPC + WebSocket + MCP. Open source under MIT.
Each source file becomes an autonomous Claude agent communicating via MCP — surfaces contract mismatches and assumption bugs through OpenTelemetry traces.
Five custom datasets across math, knowledge, and RLHF — used in #1 leaderboard models and SingleMoM expert composition experiments.
Over 10,000 documented experiments in Weights & Biases — sweeps, ablations, and training runs underpinning every published technique.
Two decades of building in public — from glFTPd community tools in C/TCL/SQL in the early 2000s, to performance-first Docker images in 2016, to neural-net debuggers, admission mutators, and smart-home IoT today.