IndexWorkAbout PlaygroundContact
Index / Playground

Ask my
portfolio anything.

A tiny runnable RAG system whose corpus is my own résumé, projects, and notes — plus a role-fit lens that re-argues the portfolio for whatever job you're hiring for.

Live RAG against this site.

§ 01 — grishman.rag
grishman.rag · local sim · pgvector mock · 8 docs indexed
›

Role-fit lens.

§ 02 — Recruiter mode

Pick the role you're hiring for and the page re-argues the portfolio against it — fit score, the strongest evidence, and where the gaps are. Designed to save you the work of squinting at a résumé wondering whether it matches your job description.

96%
Fit signal
Excellent fit
Strongest evidence
  • 4+ years shipping production GenAI: RAG pipelines, LLM orchestration, evals.
  • Owns the entire stack — from PgVector / Mongo Vector retrieval to React / Next.js front-ends.
  • Wrote prompt + retry logic that resolved real prod 429s on streaming generation.
  • Published OCR + indexing research — comfortable in both research and product surfaces.

Publications.

§ 03 — Research
Paper · Peer-reviewed

Implementing Video OCR with the SWT Technique
for Video Indexing and Analysis

Published research focused on extending traditional OCR to extract textual information from video content for efficient indexing and analysis. The system was developed in Python, integrating modules including OpenCV2, PyTesseract, and a speech-recognition library.

The pipeline involved video indexing, text segmentation and recognition via the Stroke Width Transform (SWT) + Tesseract, and parallel speech recognition using MoviePy to generate searchable digital transcripts.

PythonOpenCV2PyTesseract SWTMoviePySpeech recognition