ai / ml engineer · m.s. computer science @ uta · class of '26
i build llm systems that hold up outside a demo.
01 Currently
rock permeability ml — predicting permeability from NMR T2 measurements across 720 carbonate core samples. XGBoost, deep neural nets, nested Optuna search, SHAP attribution. Manuscript in progress at the iResearchE3 Lab, UT Arlington.
02 Selected Work
|
|
mindful-ai
agentic ai platform · voice + memory, langgraph routing across claude / gemini / llama, pgvector rag. live on vercel. |
|
|
jobagent
human-in-the-loop job-application pipeline · discovers roles, scores match, tailors a fact-verified resume. never auto-submits. |
|
|
portfolio
the whole story on one page · projects, research, publication. nanduu24.github.io |
|
|
datadoomsday
applied ml notebooks · data analysis and modeling experiments. |
|
|
esportsbot-ai
esports q&a bot · nlp over game and match data. |
|
|
fleetiq
grounded hybrid document agent for trucking fleets · llm router across sql / retrieval / both, inline citations, honest "i don't have that." team project. |
03 Instruments
|
Python |
TypeScript |
LangChain |
FastAPI |
Next.js |
React |
Postgres |
|
PyTorch |
TensorFlow |
scikit-learn |
pandas |
Docker |
Supabase |
Anthropic |
04 Published
Sign Language Caption Generation Using LSTM — IEEE International Conference for Convergence in Technology (I2CT), 2024. Real-time sign-language captioning from MediaPipe Holistic keypoints, decoded by an LSTM action-detection model. ieeexplore.ieee.org/document/10543890
05 Correspondence
06 The Receipts

