Manish Shaw

I’m Manish Shaw, currently pursuing a four-year B.S. in Economics at the Indian Institute of Technology Bombay, with a complementary minor in Artificial Intelligence & Data Science from CMiNDS. My dual focus on economics and AI equips me to blend quantitative rigor with cutting-edge technical solutions.

Last summer, I interned at Tanumanasa as an AI Developer, where I helped architect a research-paper–oriented chatbot for improved comprehension. In that role, I coordinated a cross-functional team, delegated tasks based on individual strengths, and directly pitched to clients—gathering feedback that drove rapid, user-centered iterations. I also formulated efficiency-boosting plans that cut development costs without sacrificing performance.

In December 2024, I led my Smart India Hackathon team to a national victory out of over 7,500 entrants. We built an ISL (Indian Sign Language) translation pipeline using LSTM and multi-headed attention, integrated speech-to-text and Neural Machine Translation, and rendered ASL gestures with 3D avatars via MediaPipe. I also devised an ultrasonic-signal-based system to detect sign-language users in Google Meet, ensuring they remained visible during multi-speaker sessions.

More recently, at the iDEA Hackathon in March 2025, I developed a tripartite transaction-graph algorithm for real-time fraud detection. I transformed transaction logs into continuous, d-dimensional embeddings using DeepWalk, designed inductive pooling for unseen data, and combined a Graph Neural Network with XGBoost to deliver robust risk assessments under live conditions.

Beyond competitions, I’ve tackled varied AI challenges: in a Hilti-hosted hackathon, I built a RASA-powered file-management chatbot with FAISS and FlashRank for semantic search; I designed a physics-informed neural model for Li-ion battery State-of-Health estimation, hitting an R² of 0.97; I deployed LSTM, GAN, and GRU architectures to predict stock price movements; and I contributed to the end-to-end design and flight testing of a quadcopter during our Aerodynamics course project.

On campus, I serve as Manager of the AI Community within the Institute Technical Council, mentoring junior engineers, spearheading “Insti-GPT” development, and organizing workshops. As TechFest Coordinator, I secured corporate sponsorships, launched national outreach campaigns, and liaised with over 10 companies to deliver seamless event experiences.

Technically, I’m proficient in C/C++, Python, Java, SQL, Docker, Flutter, and Git; I’ve worked extensively with PyTorch, TensorFlow, Hugging Face Transformers, LangChain, and FAISS. Certified in Docker (KodeKloud) and PostgreSQL, I’m passionate about building scalable, user-centric AI systems that drive real-world impact.