Srijit Panja
@srijitpanja
Generative AI Lead at Innovapptive
Hyderabad, India
Srijit Panja is an experienced AI and ML professional specializing in Generative AI, NLP, and building complex data solutions. He has expertise in developing RAG architectures, optimizing AI workflows, and implementing advanced models using open-source and proprietary LLMs. His experience spans e-commerce, healthcare, and industrial automation, focusing on data-driven insights and operational efficiency.
Experience
Generative AI Lead
Innovapptive
Led the enhancement of AI workflows for factory operations, improving efficiency using RAG architecture and Selective Prompt Optimization (SPO). Directed the digitization of manuals and forms, developing a two-flow form-JSON generation process using Sentence Transformer and Chroma. Developed a system to clone paper-based forms by integrating an Image+Prompt input pipeline with a Vision-LLM.
ML Engineer - Consulting
Saras Analytics
Built a Demand Forecasting Model for health supplements using non-linear regression and developed an E-commerce Media Mix Modelling (MMM) capability using Autoregressive modeling and Bayesian optimization. Led the development of a conversational chatbot (Daton Product Assistant) using RAG with Langchain and Llama-2, and developed a POC for automated data sourcing and dashboard creation using modified-RAG with schema context to VertexAI.
Data Scientist
Cognida.ai
Co-led an NLP chat-bot for a health insurance client using GPT-3.5 and a dual-database architecture (Qdrant/CouchDB). Developed an API component for an QnA pipeline using Haystack and fine-tuning DistilBERT. Developed a POC for structured data extraction from documents using DocumentAI, Tesseract OCR, and LayoutLM.
Software Engineer G2 - Data Science
Infogain
Developed an end-to-end ML solution to predict patient no-shows using XGBoost and Random Forest. Engineered a data-driven clustering solution for high-risk patient segments using K-Means and hierarchical clustering. Contributed to a self-service BI tool by generating synthetic text data using a Variational Autoencoder (VAE).
Volunteer Data Scientist
United Nations Development Programme - SDG AI Lab
Represented India in the SDG AI Lab initiative.
Education
Cochin University of Science and Technology
Master’s (M.Sc.) in Computer Science
Machine Intelligence
GPA: 8.65
University of Calcutta
Bachelor’s (B.Sc.)
Math + Phy + C.S
Score: 62.08/100