Skip to content
View vin0san's full-sized avatar
✴️
✴️

Highlights

  • Pro

Block or report vin0san

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.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

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

Report abuse
vin0san/README.md

Hi, I'm Vineet 👋

Electronics Engineering Undergrad ('22–'26) | ML Optimization & Hardware Acceleration

I'm obsessed with making ML efficient. Not just accurate- efficient. Quantization, pruning, distillation, edge inference. ECE background means I think in hardware constraints, not just model metrics.

Current Interests: ML optimization, hardware acceleration, edge inference.


🛠️ Technical Toolkit

Category Tools & Technologies
AI/ML Python PyTorch ONNX HuggingFace
Systems/Backend Go FastAPI C++
Cloud & DevOps AWS Docker Linux
Frontend/Data React PostgreSQL MongoDB

🚀 Selected Work

Sign Language Recognition

  • 98.81% validation accuracy on ASL fingerspelling, ~25 FPS on CPU
  • MediaPipe hand landmarks → PyTorch MLP → ONNX Runtime
  • Roadmap: quantization, pruning, distillation, edge deployment
  • 🔗 Code

nano-url-engine

Contract Intelligence Parser

  • 50MB+ PDFs in <10 seconds, 95% extraction accuracy
  • FastAPI, React, MongoDB, Docker
  • 🎥 Demo • 🔗 Code

Unsupervised Fraud Detection

  • 0.93 ROC-AUC, ~71% recall @ 2% flagging threshold on 280k transactions (GMM)
  • 📓 Notebook

Go Web Crawler

  • 10k+ URLs in <5 seconds, goroutines + channels
  • 🔗 Code

🌟 Beyond the Code

I approach ML optimization from first principles—paper-first, implement-from-scratch. Worked through the fundamentals (MLP → LeNet → VGG → ResNet → RNN/LSTM → Transformers → ViTs → MobileNets/EfficientNets), and now moving into ML systems theory. Understanding the architecture, not cargo-culting the code.

Picking up Japanese (JLPT N4/N5 target: Dec 2026). Long game.

Off-duty? Competitive programming, systems design problems, and reverse-engineering how things actually work.

📫 Let's Talk

LinkedIn Blog

Open to OSS collaboration in quantization/optimization toolchains. Let's talk about making ML faster and leaner.

Pinned Loading

  1. skin-lesion-onnx-api skin-lesion-onnx-api Public

    An end-to-end medical AI system featuring an EfficientNet-B3 backbone optimized via ONNX for high-speed inference, served via a FastAPI microservice and Streamlit UI.

    Jupyter Notebook 1

  2. Contract-intel Contract-intel Public

    Python

  3. retrieval-engine retrieval-engine Public

    Python

  4. url-shortener url-shortener Public

    A production-inspired URL shortener built with FastAPI, PostgreSQL and Redis.

    Python