###
🔭 Current Focus
🧱 Training AI models on real Indian construction field data (NHAI, RMC, precast, ports, railways)
📐 Building Visual SLAM pipelines for construction site progress tracking
📏 Automating quantity takeoff from architectural floor plans with AI
🎯 4+ years hands-on in India's construction & building materials industry — that's the edge I build from
🖥️ Core Languages
🧱 ConMix.ai — AI Concrete Mix Optimization
Trained on a proprietary 1,121-row Indian field dataset spanning NHAI, RMC, precast, ports, and railway projects.
- 3 live models on Hugging Face — strength prediction R² up to 0.883
- CI parity check wired into GitHub Actions
- Live Gradio demo deployed on Render
📐 TracProgress.ai — Visual SLAM Site Progress Monitoring
https://github.com/Siddartha-DevOps/tracprogress-ai Computer-vision pipeline comparing planned vs. actual construction progress from site footage.
- Built on ORB-SLAM3 + COLMAP + GTSAM
- Public demo live on Vercel
- AI defect detection module in active development
📏 TakeOff.ai — AI Construction Quantity Takeoff
https://github.com/Siddartha-DevOps/TakeOff.git AI-driven quantity takeoff from architectural floor plans.
- Custom training harness (prepare.py / train.py)
- Honest fallback UI when AI auto-detect is unavailable — no fake results shown to users
🤖 SiteGist.co — AI Chatbot Builder for Construction Businesses
🤖 SiteGist.co — AI Chatbot Builder for Construction Businesses
- Programmatic SEO + directory listing go-to-market strategy
- Live consulting clients in the construction/engineering space
I'm always excited to discuss:
- Applied AI in construction / civil engineering
- Real-world ML on messy, domain-specific field data
- Pre-seed fundraising for AI-in-construction startups
- Founder-market fit stories from non-traditional backgrounds
"🚀 From civil engineer to AI founder — building what I understand from the ground up."