Harshavardhana Srinivasan
GenAI Engineer | LLM Apps, RAG, Agents | Time Series Forecasting | Ex-AstraZeneca (5+ yrs) | Texas A&M Marketing Data Scientist | Open to Work and New Oppurtunites | Number theory, Quantum & Astrophysics Enthusiast
- Role
- Ai Consultant at Headstorm
- Location
- Dallas, TX, US
- LinkedIn followers
- 500 followers
About Harshavardhana Srinivasan
I’m a Senior Data Scientist and GenAI Engineer with 5+ years at AstraZeneca and current consulting plus research experience building LLM systems, forecasting products, and scalable analytics pipelines that drive measurable business impact.Currently at Headstorm AI (Addison, TX), I build production AI systems end to end. Most recently I built a real time Sales Coach with chat and voice guidance (AssemblyAI + Azure SSML) that reduced speech latency by 67%. I also built an intelligence layer using OCR driven persona plus weather and property signals to personalize contractor sales conversations. In parallel, I delivered a demand planning forecasting system across 205 vendor level series achieving 82% accuracy, with automated model routing and business ready dashboards (marimo).At Texas A&M (MS Data Science, GPA 4.0), as a Graduate Research Assistant I modeled compensation performance dynamics for employees using Random Effects approaches and network metrics (NetworkX). I also built an end to end RAG pipeline (LangChain + GPT 3.5) over 200+ Gong sales transcripts, reducing manual analysis by 87%(117 hours saved) and producing behavioral insights that informed survey designs across 5 companies.Previously at AstraZeneca, I built ML and automation pipelines used globally, supporting 30K+ users and saving hours annually. Highlights include TF IDF + XGBoost classification of service requests per month (saving 20+ hours per month and improving SLA compliance by 15%), A/B testing that increased email open rates by 18%, and SnapLogic/Airflow automation that reduced reporting from 5 hours to 20 minutes across 10 APAC markets.I also enjoy building in public. I built a Graph + GPT 3.5 hybrid recommender improving Precision, Recall, and NDCG at K by about 50%, and I open sourced an ORPO tuned LLaMA 8B model (230+ downloads). Open to GenAI Engineer, Applied ML, Data Science, and Forecasting or Decision Science roles.
Experience
Ai Consultant
Oct 2025 — Present · Addison, TX, US
AI consultant to SRS distribution (Home Depot)
Education
Texas A&M University
Masters in Science, Data Science
International Institute of Information Technology Bangalore
Post Graduate Diploma, Data Science
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.