Abhinav Kumar

Data Engineer (AI/ML) at Inventia

Designed and deployed production-grade AI/ML backends using FastAPI and OpenAI API for analytics, chatbot, and insight-generation workflows across multiple products. Built ML pipelines for electricity utility analytics covering forecasting, anomaly detection, voltage diagnostics, and solar consumption prediction using XGBoost, LightGBM, TensorFlow, PySpark, and Dask. Developed a multi-product AI support chatbot with hybrid RAG (PostgreSQL lexical + Qdrant vector search), Redis session management, and module-level permission-based knowledge retrieval. Built an intent classification and routing layer combining LLM-based classification (OpenAI/Ollama) with regex fast-paths, Redis caching, and async downstream API orchestration. Developed an AI Insights engine for dashboards that generates contextual KPI and chart insights using OpenAI API, with per-product configurable priorities stored in a database.