I build applied AI and ML systems in New York, NY — LLM applications, models, and the data pipelines and automation behind them. Neuroscience is what got me here, and why I care about knowing why a model does what it does.
🌐 myesmin.github.io — project write-ups, full experience and résumé · 💼 LinkedIn · ✉️ myesmin415@gmail.com
- 🔬 Research Assistant at DAIR Lab: graph neural networks for protein research
- ☸️ Kubeflow org member: Python and Go contributions to the Model Registry and SDK (hub, sdk)
- 📖 Building Sera: a reading companion that reads a PDF aloud with word-synced highlighting — FastAPI + React/TypeScript, Postgres with pgvector, OCR, a queued TTS worker and streaming LLM answers, running on AWS (currently on voice intelligence)
- 🛡️ Building ModSieve: a first-pass moderation tool that auto-resolves the Reddit posts it's sure about (22.5% at 99.2% precision) and sends the rest to a human
- 🏢 Energy Data Manager at EN-POWER Group: emissions models and LL97 compliance work for NYC buildings
| Languages |        |
|---|---|
| ML & AI |           |
| Data |          |
| Backend & tools |          |
| Productivity |     |
- I studied Neuroscience & Behavior at Wesleyan, and the link between how brains learn and how models learn is what pulled me into ML.
- I like turning many messy data sources into one answer I can defend, and making analysis useful to people who never open a notebook.
- Open to roles in applied AI / ML / data / software engineering.