Electrical Engineer with CS Specialization. Building Full Stack Systems and applications along with data-oriented projects, AI applications, Open source contributions and basically everything that tires me out...
- 📬 Reach me at: punyak.dei@gmail.com
Note: Some of these projects are currently private or in active development until their official release
I actively contribute to popular open source repositories. Here are some of my recent contributions:
- scikit-learn/scikit-learn: PR #34817 - Optimized
confusion_matrix_at_thresholdswith integer cumulative sum for unweighted inputs. - scikit-learn/scikit-learn: PR #34789 - Preserved pandas DataFrame inputs and feature names across RFE and RFECV.
- explosion/projects (spaCy): PR #211 Replaced OS-specific bash commands with cross-platform inline Python.
|
URL shortener: FastAPI + ReactJS, JWT auth with sessionStorage, DynamoDB, bijective short-code generation, and 302 redirection. Deployed on AWS EC2 (t3.micro) with IAM role auth, elastic IP, and Caddy SSL. Gist: System Design |
Open-source Retrieval-Augmented Generation (RAG) system with persistent vector storage (ChromaDB), multi-provider LLM orchestration (LiteLLM), and strict source citations. Built with Flask backend. |
|
Full-stack deployment of sentiment classifiers in Scikit-learn and JAX frameworks. Inferences handled via FastAPI on a Railway backend, connected to a vanilla JS frontend on Vercel. |
Structured a robust research workspace focusing on multimodal product price prediction from catalog text descriptions and images. Benchmarked TF-IDF + LightGBM to BERT fine-tuning regressors. |
|
Implemented a tabular Q-Learning agent for Tic-Tac-Toe. Fully reproducible training using configurable hyperparameters and random seeds via config.yaml. |
| Category | Technologies |
|---|---|
| Languages |       |
| Backend |   |
| Deployment |     |
| Machine Learning & AI |     |
| Data Eng & Storage |      |