Enhancing AI Logistics & Disaster-Resilience Platform

Project Title NER-SMART: AI-Powered Logistics & Disaster-Resilience Platform for Northeast India (Streamlit) Project Overview NER-SMART is a Streamlit-based logistics intelligence platform built for the Northeast Indian Region (NER) — a hilly, disaster-prone corridor where landslides, flooding, and road closures routinely disrupt supply chains. The app models the region's road network as a live graph, blends weather + terrain + incident data through a machine-learning risk model, and gives field teams a single dashboard to plan routes, track vehicles, log incidents, and simulate disaster scenarios before they happen. The core app is built and functional. I'm looking for a freelancer to help polish, debug, and extend it. Key Features (already built) GIS Map — interactive map of NER districts and highway network AI Risk Prediction — a RandomForest model (scikit-learn) that scores road-segment risk from rainfall, slope, traffic, and incident history, with a transparent contribution breakdown Route Optimizer — real road routing via OSRM, comparing AI-recommended / fastest / lowest-risk paths with live ETA and risk scoring Fleet Tracking — live map of vehicles in transit, with GPS trail history and a documented JSON API contract for connecting real GPS hardware Field Reports — geo-tagged incident reporting (landslide, flood, road damage, etc.) that instantly feeds back into the live risk model Disaster Simulator — adjustable sliders (rainfall multiplier, landslide bias, road-condition penalty) to stress-test the network under hypothetical scenarios AI Copilot — a rule-based natural-language Q&A interface over the live network state (road risk, blocked roads, safest routes, at-risk vehicles, open incidents) Tech Stack Python · Streamlit · scikit-learn · NetworkX · Folium / streamlit-folium · Pandas/NumPy · OSRM routing API · Streamlit Community Cloud Scope of Work (customize as needed) Fix outstanding bugs (map rendering, route calculation edge cases, layout issues) Resolve deployment issues on Streamlit Community Cloud (Python version pinning, dependency conflicts) Improve the AI Copilot with real LLM integration (Anthropic/OpenAI API) instead of the current rule-based fallback Add authentication / multi-user support Connect a real GPS/IoT data source to Fleet Tracking General UI/UX polish and mobile responsiveness Code review and refactor for production readiness Ideal Candidate Strong Python + Streamlit experience Familiarity with geospatial tools (Folium, OSRM, or similar routing APIs) Comfortable with scikit-learn / basic ML model integration Bonus: experience with disaster-response or logistics-domain applications Deliverables Working, deployed Streamlit app (Community Cloud or platform of your choice) Clean, commented, version-controlled code (GitHub) Brief documentation of any changes made Timeline & Budget (fill in your specifics — e.g. "1–2 weeks, fixed price" or "ongoing, hourly")