About

I’m Prathamesh Devadiga, an AI researcher, engineer, and educator pursuing a B.Tech in Computer Science at PES University, Bengaluru. My focus lies at the intersection of Machine Learning Systems (MLSys), AI Security, and Deep Learning. I’m passionate about building reliable, scalable, and interpretable AI systems rooted in real-world impact and open-source philosophy.

Currently, I’m contributing to Google Summer of Code 2025 with UC Santa Cruz, where I’m building billion-scale high-dimensional vector embedding datasets to benchmark Approximate Nearest Neighbor (ANN) algorithms. I also conduct research on Trustworthy AI and Security at Dartmouth College, focusing on adversarial robustness and secure ML pipelines.

What I Do Link to heading

Research and Development
I lead Ādhāra AI Labs, an independent research lab focused on Generative AI, LLM systems, and Retrieval-Augmented Generation (RAG). We design open-source tools that help bridge the gap between cutting-edge research and production-ready applications.

Past research work includes:

  • Deep learning for PDF malware detection at IIT Indore
  • Applied ML for Social Media research at Deakin University

Teaching and Mentorship

  • Co-instructor for a 30-hour “Deep Learning from Scratch” course at PESU IO
  • Teaching Assistant for ML courses at PES University
  • Mentor at Women in Data Science (WiDS), Google Developer Student Clubs, and several AI/ML hackathons

Engineering and Open Source

  • Built generative AI workflows at a stealth AI startup
  • Developed a RAG-based Go tool for code review and optimization

Education & Learning Link to heading

  • Oxford Machine Learning Summer School – Explored advanced topics in generative AI, statistical ML, healthcare ML, and geometric deep learning
  • Cohere ML Summer School – Participated in hands-on workshops focused on large language models, prompt engineering, and deployment of NLP systems

Initiatives Link to heading

  • Ādhāra AI Labs – Open-source AI research lab focused on GenAI and LLM infrastructure
  • Catalysis – A nonprofit initiative democratizing STEM education through mentorship
  • The Bookshelf – Peer-to-peer platform for student book lending and academic collaboration

Skills and Interests Link to heading

Core Areas: Generative AI, Retrieval-Augmented Generation (RAG), Machine Learning Security, Explainable AI (xAI), ML Systems, Deep Learning, Natural Language Processing (NLP)
Frameworks & Libraries: PyTorch, TensorFlow, HuggingFace Transformers, LangChain, Scikit-learn
Programming Languages: Python, Go, Bash
Tools & Platforms: Weights & Biases, MLflow, Docker, Git, FastAPI, Jupyter, REST APIs
Research Interests: Secure and interpretable ML systems, scalable GenAI workflows, efficient training and inference, adversarial robustness, applied LLMs