Machine Learning Engineer | AI Engineer | Software Engineer | Data Scientist
I’ve worn different hats throughout my career — I build production-grade AI systems end-to-end: data → models → infra → developer experience.
- 🌎 San Francisco Bay Area
- 🧠 Focus: LLM systems, RAG, multimodal AI, scalable backend/data platforms
- 🤝 Open to: OSS + applied AI engineering collaborations
- GitHub: https://github.com/shashidharbabu
- LinkedIn: https://www.linkedin.com/in/p-v-d-shashidhar-babu-9b91471b9/
rocketride-org/rocketride-server
High-performance AI/data processing engine with 50+ pipeline nodes, multi-agent workflows, and SDKs (TypeScript, Python, MCP).
- RocketRide Server — production AI pipeline engine (C++/Python, SDKs, Docker)
https://github.com/rocketride-org/rocketride-server - Aerive Platform — microservices travel booking platform (React/Vite/Tailwind, API gateway, Kafka, Redis, MongoDB, Postgres)
https://github.com/shashidharbabu/Aerive-Platform - Multimodal RAG — multimodal retrieval system (text/image/audio/video)
https://github.com/shashidharbabu/multimodal-rag - Transformer from scratch — end-to-end transformer implementation
https://github.com/shashidharbabu/transformer-architecture - AI NutriCoach — food-image insights + personalized nutrition ideas
https://github.com/shashidharbabu/ai-nutricoach
“data, models, evals & robots(new interest!)” — I care about the full system: correctness, performance, infra, and shipping.
🧪 How the art at the top is made
Everything above is self-contained animated SVG — no third-party stats service, no
GitHub token, no JavaScript. GitHub strips <script> from READMEs but does render
SVGs via <img> and runs their CSS keyframe animations, so all the motion lives
inside the SVG files themselves.
| File | Made by | Refreshed |
|---|---|---|
contrib-heatmap.svg |
scripts/fetch_contributions.py → scripts/render_heatmap_svg.py |
daily, by GitHub Actions |
avi-ascii.svg |
scripts/prep_photo.py → scripts/make_ascii_svg.py |
manually, when the photo changes |
info-card.svg |
scripts/make_info_card.py |
manually, when the details change |
Contribution data is scraped from github.com/users/<user>/contributions — the
same public HTML fragment the profile page itself uses, so no auth is required.
.github/workflows/update-profile-art.yml re-scrapes and re-renders it on a daily
cron and commits the result.
Regenerate locally:
python -m venv .venv && source .venv/bin/activate
pip install -r scripts/requirements.txt # heatmap
pip install -r scripts/requirements-portrait.txt # portrait
# heatmap (this is what the daily Action runs)
python scripts/fetch_contributions.py
python scripts/render_heatmap_svg.py
# portrait - the crop frames the subject out of a wider full-body shot, and
# --aspect is tuned so the portrait renders the same height as the info card
# (re-derive it whenever you add or remove rows from the card)
python scripts/prep_photo.py source-photo.jpg \
--crop 0.214,0.055,0.786,0.663 --aspect 1.594
python scripts/make_ascii_svg.py
# info card
python scripts/make_info_card.pySet STATIC=1 on any of the render scripts to emit a frozen frame instead of an
animated one — useful for local previews.


