I’m Harshvardhan Singh, an AI Engineer who loves turning messy real-world problems into clean, reliable software. I enjoy working at the intersection of machine learning + backend systems, especially when the end result is a product people actually use.
- 🔭 Interested in Applied ML, Computer Vision, LLMs/Agents, and scalable backend engineering
- 🧩 I like building end-to-end: model → API → product → deployment
- 🎯 I care a lot about performance, reliability, and developer experience
- 🏗️ Building an Exercise Form Evaluation App using MediaPipe + OpenCV + FastAPI (pose tracking → joint angles → actionable feedback)
- 📚 Learning more about LLM tooling, RAG, and agentic workflows
- ⚙️ Improving my distributed systems and deployment skills (Docker, cloud services, CI/CD)
I’m open to collaborating on:
- 🧠 AI product engineering (LLM apps, RAG, agent workflows)
- 👁️ Computer vision projects (pose estimation, OCR, document AI)
- 🛠️ Backend systems (APIs, pipelines, data stores, performance tuning)
If you’re building something interesting and need an extra pair of hands, feel free to reach out!
Swap the links below with your actual repos / demos.
-
🧠 Hybrid ML Scheduler (CPU–GPU)
RL-based scheduler that routes ML tasks across CPU/GPU using FastAPI + Redis + WebSockets to reduce makespan.
🔗 Repo: https://github.com/harshvardhan579/hybrid-ml-scheduler -
🏋️ Exercise Form Evaluation App
Computer vision app that analyzes workout videos to compute joint angles and deliver form feedback via an API.
🔗 Repo: https://github.com/harshvardhan579/exercise-form-evaluation
- 💼 LinkedIn: https://linkedin.com/in/harshvardhan2
- 📧 Email: harshvardhan6970@gmail.com
- 🌐 Portfolio: https://hvs-tau.vercel.app/
I enjoy building projects where AI meets real users, and I’m always down to ship something end-to-end rather than stopping at a notebook.
