Projects

Problem: Neuroscience teams need to inspect calcium-trace events quickly without forcing every experiment through a slow desktop workflow.

Approach: Browser-based analysis interface in Next.js and TypeScript that makes neural traces explorable, shareable, and easy to pair with downstream event-detection research.

Result: The surrounding research workflow reached a 0.92 F1 event-detection score. Ships as a live demo for interactive inspection instead of static plots.

Why it stands out: Combines research credibility with product execution: live UI, domain-specific data exploration, and measurable modeling impact.

Problem: Long expert-call transcript archives are difficult to search, summarize, and cite reliably when every answer has to stay grounded in source material.

Approach: Local RAG stack with Llama 3.1, LlamaIndex, LanceDB, and a Gradio interface — retrieval, inference, and source citation all stay modular and replaceable.

Result: Runs fully locally, returns source-grounded answers over transcript collections. Documents roughly 1–2 minute response times on local hardware.

Why it stands out: Shows applied retrieval depth rather than just prompting: local inference, modular retrieval design, and explicit source grounding.

Problem: Content publishing breaks down when image selection, captioning, search, and posting still require manual handoffs across tools.

Approach: Serverless AWS pipeline connecting image retrieval, Rekognition analysis, Claude-based captioning, OpenSearch indexing, and publishing logic into one automated flow.

Result: Automates a 5-stage content pipeline end to end, reducing the workflow to one orchestrated system instead of separate manual steps.

Why it stands out: Strong systems work: event-driven architecture, model orchestration, and production-minded automation instead of a single model demo.

Problem: Collections teams need a usable operating surface for turning messy account data into actionable outreach plans rather than static spreadsheets.

Approach: Full-stack product with FastAPI, React, SQLite, and the OpenAI API to generate strategies, manage contacts, and organize actions in a timeline-based UI.

Result: Handles both Excel and PDF imports, then turns data into AI-generated multi-step strategy blocks inside a single workflow interface.

Why it stands out: Clearest product build in the portfolio: ingestion, reasoning, UI state, and decision support packaged as one cohesive application.