I’m Pranav, an AI engineer who builds RAG assistants, LLM agents and predictive models on real plant data, most recently at Mercedes-Benz and Veeco.
Engineers and technicians using the plant assistant
Answer acceptance on the RAG assistant
Unplanned downtime on pilot lines
Earlier drift warning than static alarms
Plain-English questions, cited answers from SOPs, maintenance manuals and quality reports in seconds.
Agents that turn requests into SQL, refresh dashboards, and raise alerts or tickets, with a person approving each action.
Anomaly detection and failure prediction on sensor time series, wired to alerts and work orders.
MES, SAP, quality inspection and line-sensor data in one Azure Databricks lakehouse.
A decoder-only transformer in NumPy and core Python: multi-head attention, layer norm and greedy decoding, trained on email to draft replies and summaries.
Multi-label NLP model on the Jigsaw dataset that flags six kinds of toxicity, served through a Flask API in Docker.
A dependency-free static site generator in Node that builds every page from one content file, with CI checks and deploys through GitHub Actions.
