Akhil Vishnubhotla

I’m a BS student in Mathematics and Computer Science with hands-on experience building production-grade systems across software engineering and applied machine learning. I’m drawn to work where deterministic logic meets data-driven modeling, and where performance, reliability, and latency matter.

As a Software Engineer Intern at MetaFloor.ai, I built a net-new production readiness engine on top of an existing ERP, implementing Sales Order ingestion and recursive BOM explosion across a 3-level hierarchy spanning 6+ ERPNext doctypes. I designed a deterministic supply-readiness service that fuses scrap/yield-adjusted BOM explosion, multi-source demand aggregation, and a recency-weighted supplier confidence model keyed to 90-day defect recurrence—benchmarking at p95=38ms for confidence scoring and p95=5ms for end-to-end readiness checks.

As a Machine Learning Engineer Intern at Biomet.life, I built a GHG emissions forecasting model for automotive manufacturers by combining historical emissions, financial/production data, and GPT-4o sentiment analysis of sustainability disclosures—improving out-of-sample R² from 0.40 to 0.76 using NLP-derived polarity features. My projects include MedCLIP-TB (CLIP-style multimodal contrastive learning for zero-shot TB classification) and a Financial News Sentiment Trading Backtester built with FastAPI + background workers and FinBERT-driven inference.