I build production backends and AI systems for teams that need to ship β not demos that look good in tabs.
CS @ KIIT '27 Β· Currently at CWI Studios Β· Bokaro / Bhubaneswar.
| What it is | Why it matters | Stack | Live | |
|---|---|---|---|---|
| π IndustryERP | Production ERP at a steel fabrication business | 7+ employees use it daily. 172 endpoints. Replaced their spreadsheets. | FastAPI Β· PostgreSQL Β· Cloud Run | β |
| π‘ OneAlert | OT/ICS cybersecurity SaaS β AI remediation, SBOM, compliance | 26β Β· 166 tests Β· 10 shipped phases Β· MIT licensed | FastAPI Β· React Β· Cloud Run | β |
| π€ TaskForge | Multi-agent crisis logistics system | Built for Google Cloud Gen AI Academy APAC 2026 (Cohort 1) | Gemini 2.5 Β· MCP Β· pgvector | β |
| π Redline AI | AI emergency dispatch IVR (fork) | Contributed Gemini integration + CI/CD + production deployment | FastAPI Β· Twilio Β· Whisper | β |
| π SkyDash | 20Hz MAVLink drone telemetry dashboard | Real-time streaming, backpressure handling | FastAPI Β· React Β· WebSockets | β |
I treat the shop floor and the customer call as primary sources, not the GitHub trending page. Before writing the ERP, I spent weeks interviewing operators at a steel plant. That's why the system actually gets used β and why I reach for MCP, agentic AI, or real-time streaming only when they solve a real problem.
Python Β· FastAPI Β· PostgreSQL Β· Redis Β· Docker Β· GCP Cloud Run Β· Cloud Build Β· GitHub Actions Β· Prometheus Β· Grafana
Comfortable with MCP, agentic systems, RAG, pgvector, and real-time streaming pipelines when the problem calls for them. RHEL-certified. GitHub Developer Program member.
π Building Blue β internal platform powering CWI Studios' operations π± Open to backend / systems / AI engineering internships where I can ship to real users from day one
π« anshajkumar341@gmail.com Β· LinkedIn Β· Portfolio




