Headline

Data Scientist → ML & AI Engineer | Building Production RAG Systems, Forecasting Models & Multi-Agent LLM Apps | Python · LangChain · FastAPI

About

Most ML engineers can train a model. Fewer can take it to production. In 2.7 years I've done both — built forecasting systems predicting revenue across multiple countries, automated healthcare data pipelines that cut processing time by 80%, and shipped RAG and multi-agent AI applications that actually run outside of a Jupyter notebook. I work across two lanes: Traditional ML — time-series forecasting with Prophet, classification and clustering with XGBoost and Random Forest, feature engineering, and model optimization. Real projects: fraud detection, customer segmentation, demand forecasting. Generative AI & LLMs — RAG systems with LangChain + FAISS, multi-agent architectures with LangGraph and CrewAI, semantic search, vector database optimization. Real projects: document Q&A systems. Most of the time these two lanes overlap. That's where I do my best work. Tech I work with daily: Python · XGBoost · Prophet · LangChain · LangGraph · FAISS · ChromaDB · FastAPI · Docker · PostgreSQL · MongoDB What I'm looking for: A team building production ML or AI systems that solves real problems — forecasting, recommendations, RAG pipelines, agentic AI, or all of the above. I'm not looking for a notebook job. Currently going deep on LLMOps and production deployment patterns. 📂 See what I build: github.com/Jayraj-Mehta 📩 Open to opportunities — let's talk.

Location

Surat

Work experience

Analyst 2 - Data Science at Accordion, Jul 2023 - Feb 2026 · 2 yrs 8 mos, Hyderabad, • Built and deployed a multi-country revenue forecasting system using Prophet, integrating Google Trends and macroeconomic indicators with adaptive model selection logic that dynamically optimized for MAPE — enabling reliable predictions across varying market conditions at scale • Automated large-scale healthcare financial data processing with Python, eliminating manual calculations and reducing per-record processing time by 80% across production workflows • Engineered end-to-end data transformation pipelines for healthcare financial datasets, implementing validation logic and error handling that ensured data quality and zero-defect processing reliability in production • Redesigned and migrated automation workflows to a new platform, doubling weekly processing capacity through code optimization and architectural restructuring while maintaining full operational continuity during transition • Partnered with business stakeholders to translate complex analytical requirements into production-ready solutions, consistently delivering on time with measurable operational impact

Manager at Mettle, NIT Trichy, Jan 2021 - Jan 2022 · 1 yr 1 mo

Education

Bachelor of Technology - BTech, Metallurgical and Materials Engineering, National Institute of Technology, Tiruchirappalli

Smt. L. P. SAVANI VIDHYABHAVAN ADAJAN