Bridging the gap between low-level C++ systems and high-stakes agentic AI deployments.
- π Education: M.Tech CSE (Spec. in Computational & Data Science) @ VIT Bhopal (2027).
- π°οΈ Ex-ISRO Researcher: Built PyTorch/TF LSTM telemetry models for PSLV missions, cutting manual review by 40%.
- π AI Impact: Boosted Google Gemini instruction accuracy by 15% via structured RLHF; logged 500+ hours evaluating production LLMs.
- π Top 0.24%: Intel AI4Youth National Finalist (out of 52,000); NLP project endorsed by NITI Aayog.
- π― Currently Building: Deterministic multi-agent LangGraph orchestrators & bio-inspired State-Space Models.
| Project / Role | Tech Stack | Measurable Impact |
|---|---|---|
| Project AETHER (AI Geospatial) | PyTorch GANs GDAL Rasterio |
Engineered a self-supervised MAE + custom GAN, improving SNR in lunar Permanently Shadowed Regions by +86 dB. |
| Synapse.cpp (Low-Level AI) | C++17 RAII Adam/SGD |
Built a full forward/backward pass neural network engine entirely from scratch with zero external dependencies. |
| Nexus Scout (Agentic AI) | LangGraph Llama-3 Pydantic |
Designed a deterministic multi-agent HR pipeline with a weighted hybrid scoring engine for auto-generated outreach. |
| AI Engineering Intern | SpaCy NLP Streamlit Python |
Deployed an end-to-end AI parsing platform, slashing resume-screening time by 80% with 90%+ accuracy. |
View Full Portfolio & Resume

