prakhar shukla

I’m a software engineer focused on turning messy data flows into reliable pipelines, decision-ready analytics, and user-facing products. Most recently, I built an end-to-end async ETL pipeline extracting from multiple SAP OData APIs and loading 500K–2M billing records into PostgreSQL, alongside a 15-table warehouse design for real-time inventory intelligence.

I also develop ML-driven experiences. I built a hierarchical recommendation model delivering 80,000+ personalized part predictions across 16,000+ vehicles, and used EDA to surface business-impacting signals like reactive purchase orders and stale PO rates to drive proactive procurement.

On the product side, I ship full-stack systems that feel fast and actionable—like a Next.js TypeScript BI dashboard with real-time stock alerts, deep drill-downs, and a Gemini-powered AI chat agent with 30+ tool handlers for SQL execution, inventory lookups, and recommendations. I’m equally comfortable orchestrating the whole lifecycle with n8n (daily ETL, drift detection, health reporting, and monthly retraining), and building operational tooling—from smart dispatch systems to scalable IoT nodes—when the problem demands it.