Priya Chandrasekaran

About

I build forecasting models for production and commodity data, mostly for small energy operators who outgrew a spreadsheet but do not have an in-house data team. Recent work: a decline-curve forecast feeding a weekly dashboard, and a RAG system over years of regulatory filings so engineers stop re-reading old reports to answer basic questions. I use Python for the modeling and Supabase or Pinecone for the vector side, with OpenAI or Claude on top. I am careful about what I let a model claim confidently versus what needs a human to check. Remote, based in Calgary.