▮ File / Personal
h/albihn
builds the
harness.
Models are commoditizing. The differentiated work is the harness — retrieval, ticket graphs, ensemble routing, SPC on the outside of the model.
At Strange Loop Labs — forward-deployed AI, built by Amazon Alexa engineers — I own technical architecture and delivery across client engagements. Exec discovery through production, regulated enterprise document workflows at scale.
Internally: high-throughput agentic coding. Classical ML applied to the code-generation loop — ensemble methods, online learning, statistical quality control. Map-reduce-style ticket-driven development with ticket-rs as the persistent context graph.
I write Vanishing Gradients from the forward-deployed trenches. No theory without receipts; no claim without code.
| No. | System | Description | Go |
|---|---|---|---|
| 001 | ticket-rs | AI-native issue tracking. Git-backed, local-first, PageRank-prioritized. MCP servers for Claude Code, Cursor, Codex, Copilot, Cline, Gemini, Windsurf, Zed, JetBrains. | |
| 002 | sgrep | Grep by meaning, not keywords. 8M-parameter Model2Vec via Candle, pure Rust, sub-10ms semantic search. One static binary, zero API calls. | |
| 003 | sqlgenie | Natural-language query engine across 23+ SQL dialects. ~200× lower per-query cost than manual baseline. Inbound from TechStars, multiple VCs, and two acquirers in the first 90 days. |
| 001 | Level Up Coding | 3× faster file conversion with DuckDB. | Data engineering |
| 002 | Level Up Coding | The modern data stack is dead. | Thesis |
| 003 | Level Up Coding | Dipping your toes into JavaScript. | Cross-language |
| 004 | AI in Plain English | What the heck is a classifier? | Primer |
| 005 | AI in Plain English | Decision boundaries. | ML foundations |
| 006 | AI in Plain English | On text similarity search. | Retrieval |
| 007 | AI in Plain English | Detect anything in text — GLiNER for zero-shot NER. | NER |
| 008 | AI in Plain English | Efficient text classification with GLiClass. | Classification |
| 009 | AI in Plain English | The Qwen 2.5 model family. | LLM survey |
| 010 | AI in Plain English | GPT anywhere — running LLMs locally with Ollama. | Local LLMs |
Economics → data science → ML science → applied AI → forward-deployed engineering.
Before Strange Loop Labs: Principal AI Scientist at a stealth AR/VR startup, shipping a real-time manufacturing copilot on Meta Quest 3 and Apple Vision Pro.
Before that: owner of the size-recommendation engine at True Fit — the company's flagship product, serving millions of shoppers daily across 30K+ brands including Nike, Target, Walmart.
Each transition sharpened the same edge: pattern recognition across markets and systems.