Krishnakumar Vijayasankar

Full-Stack & AI Engineer | Ex-Ford | Building real products with PyTorch, Spring Boot & Cloud

🚀 Featured Projects

Pose Director – AI-Powered iOS App (Leaded Development)

• Try it out: [https://apps.apple.com/us/app/pose-director/id6737916480]
• Real-time pose estimation app with on-device inference and cloud backend
• Built end-to-end: model optimization, iOS frontend, API integration, and App Store deployment
• Tech: PyTorch, Core ML / TensorFlow Lite, Swift, Cloud architecture

DreaMS Chemistry Embeddings – Food Metabolomics Research

• Fine-tuned DreaMS self-supervised transformer on NDM & MassSpecGym datasets
• Generated 1024-dimensional chemistry-specific embeddings + cosine-similarity pipeline
• Improved spectral annotation accuracy for unknown metabolites in food matrices
• Tech: PyTorch, Transformers, FAISS, Mass Spectrometry data

AI Neuroscience Project – MIT Hackathons

• Developed PyTorch deep learning pipeline for micro-ECoG and EEG signal processing
• Applied digital signal processing techniques on neural data
• Tech: PyTorch, DSP, EEG/ECoG analysis

3D Psychological Horror Game – Global Game Jam 2026 (Boston)

• Built and published a complete 3D psychological horror game in 48 hours
• Handled gameplay mechanics, 3D assets, and deployment
• Tech: Godot Engine
• Play here: [Itch.io Link]

---

Tech Stack

Languages: Python, Java, JavaScript, SQL, C++
ML/DL: PyTorch, Hugging Face, Transformers, Computer Vision, Signal Processing
Backend: Spring Boot, FastAPI, REST APIs, Docker, Google Cloud Run
Frontend: Angular, Swift
Data: Pandas, NumPy, FAISS, MySQL, MongoDB

Currently pursuing M.S. in Applied Artificial Intelligence at Northeastern University (GPA 3.8).

Open to full-time roles in AI Engineering, Full-Stack Development, and ML infrastructure.