Skip to content
View as567-code's full-sized avatar
🏙️
Working from Texas
🏙️
Working from Texas

Block or report as567-code

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
as567-code/README.md

Aditya Swaroop

MS Computer Science @ UT Dallas (3.8 GPA) | Software Engineering Fellow @ Princeton Research Computing | NeurIPS 2025 Under Review

About

Building ML systems and backend infrastructure. Princeton HPC fellowship (CUDA/C++), NeurIPS 2025 submission on on-device LLM quantization, 2 published papers. Open source contributor to Celery, IPython, Apache Airflow.

Featured Work

  • ATQ-LLM — Adaptive Ternary Quantization for LLM compression (~16x on GPT-2). NeurIPS 2025.
  • MusicSourceClassifier — AI vs Human music detector. 99.9% accuracy on 900 samples. CNN + FAISS similarity search.
  • GPS Fleet Telemetry — Event-driven pipeline with anomaly detection. 43-test suite.
  • DoctorBot-AI — Full-stack AI chatbot. FastAPI + Mistral + Three.js. 18 async tests.
  • N-Body Optimization — CUDA/OpenMP/MIC. 10.4x speedup. Princeton fellowship.
  • Video Understanding — Multimodal pipeline. Qwen2-VL + BLIP + Flan-T5.

Publications

  • Adaptive Ternary Quantization for On-Device LLM Compression — NeurIPS 2025 (Under Review)
  • Quantum Enhanced Recommendation AI Chatbot — IGI Global, 2025
  • Ensemble ML for Multi-Disease Detection — EasyChair, 2023

Tech

Python · C/C++ · CUDA · TypeScript · PyTorch · Transformers · FAISS · FastAPI · React · Docker · Kubernetes · AWS · PostgreSQL

Links

Portfolio · LinkedIn · adityaswaroop23.rrps@gmail.com

Popular repositories Loading

  1. Video-Understanding-Chat-System-MVP- Video-Understanding-Chat-System-MVP- Public

    Multimodal video analysis — Qwen2-VL + BLIP frame captioning, Flan-T5 conversational Q&A over timestamped captions. Streamlit UI.

    Python 2

  2. ATQ-LLM ATQ-LLM Public

    Adaptive Ternary Quantization for On-Device LLM Compression — NeurIPS 2025 (Under Review). ~16x compression on GPT-2 via ternary weights, STE, mixed precision, QAT. 32 unit tests.

    Jupyter Notebook 2

  3. MusicSourceClassifier MusicSourceClassifier Public

    AI vs Human music detector — 4-layer CNN (421K params) achieving 99.9% accuracy on GTZAN+SONICS. FAISS similarity search with TorchOpenL3 embeddings. React + Flask full-stack.

    Jupyter Notebook 1

  4. DoctorBot-AI DoctorBot-AI Public

    Doctor Who AI chatbot — FastAPI + Mistral AI backend with 18 async tests, robust JSON parsing. React frontend with Three.js vortex (600 particles).

    TypeScript 1

  5. PythonDice-Roll PythonDice-Roll Public archive

    Python

  6. jaydeep jaydeep Public archive

    Python