About
Data Scientist with 5+ years of experience, specializing in generative AI systems such as RAG pipelines, retrieval based matching, and conversational AI, alongside machine learning, predictive analytics, and time series forecasting. Proven ability to take models from prototype to reliable, scalable production systems.
Skills & Expertise (24)
Work Experience
Staff Data Scientist
Hypersonix Inc.
May 2026 - Present
Developed a system that matches client products to their closest competitor equivalents using embedding based semantic search with attribute filtering to narrow candidates, enabling better pricing decisions. Implemented the retrieval stack with LlamaIndex and Milvus for fast, accurate nearest match search over the enriched representations, achieving 85% match accuracy. Designed a config driven architecture that adapts to each client’s requirements and scaled it across 90+ clients deployed on Databricks.
Senior Data Scientist
Hypersonix Inc.
Apr 2023 - Apr 2026
Developed a system that categorizes competitor products against a client’s own product taxonomy, giving clients a directly comparable view of competitor assortment. Mapped each competitor product to the closest tenant category using embedding based semantic similarity on Milvus. Scaled the categorization pipeline across 50+ clients on Databricks through a config driven design. Developed a conversational AI assistant that answers user queries across multiple modules such as forecasting, inventory, and more from a single interface, reducing the steps users needed to get answers. Enabled natural follow up questions by building the workflow in LangGraph with checkpoint based memory to retain conversation context. Developed a solution that measures each campaign’s real sales impact, separating high lift from low lift promotions. Designed it to scale across multiple clients, applying consistently to different product ranges and promotion types.
Data Scientist II
Hypersonix Inc.
Jul 2022 - Mar 2023
Developed a system to predict stockout and overstock incidents at the item level before they happen. Combined demand forecasts and price elasticity as model inputs to predict stockout and overstock risk per item.
Data Scientist I
Hypersonix Inc.
Nov 2021 - Jul 2022
Developed item level demand forecasting models using time series methods to predict demand at a per item level. Segmented time series by pattern type and routed each segment to the best suited model, improving weighted forecast accuracy from 68% to 81% over a single baseline model. Scaled the solution across 20+ clients, delivering reliable forecasts that adapt to each client’s data.
Data Science Intern
Hypersonix Inc.
Aug 2021 - Nov 2021
Built item level price elasticity models using regression based techniques to quantify demand sensitivity across products, enabling more precise, data driven pricing decisions.
Education
Master of Technology, Aerospace Engineering - Indian Institute of Technology, Kharagpur (IIT Kharagpur)
2019 - 2021 · Afghanistan
Bachelor of Technology, Aeronautical Engineering - Hindustan Institute of Technology and Science, Chennai
2014 - 2018 · Afghanistan
Certifications
No certifications added yet
Interested in this developer?
Profile Score Breakdown
Profile Overview
Availability Details
Visa Status
Citizen
Relocation
Open to Relocation
Skills (24)
Click a skill to find developers with the same skill
