CHANDERVEER SINGH CHAUHAN

• Developed an AI-powered full-stack environmental intelligence platform using FastAPI, React 19 (50+ components), PostgreSQL, integrating Sentinel-2 imagery, YOLOv8 land classification, 10-year plantation & economic planning models, and a Gemini-powered RAG architecture (LangChain, Pinecone, Weaviate); engineered high-performance geospatial systems with Leaflet (Texas-bound maps), Supercluster (10K+ markers), 60 FPS canvas rendering, Cloudinary, Framer Motion, and built real-time services including a Gemini 2.0 Flash citizen chatbot (WebSocket/HTTP streaming, SQLAlchemy persistence, citation grounding), NASA FIRMS fire tracking (, plus forest monitoring pipelines using Global Forest Watch GLAD alerts, Sentinel Hub, Planet Labs, Gemini 2.5 Pro multimodal segmentation, GeoJSON ETL (50K+ features), AWS S3 storage, county-level carbon calculator and automated PDF reporting.

• Architected an agentic AI-driven financial intelligence platform using Django REST Framework with 8 autonomous AI agents (earnings transcript analysis, brokerage parsing, financial modeling, intelligent comparison, chatbot) powered by Gemini, enabling real-time earnings call processing and automated extraction of financial KPIs (EBITDA, ROE, revenue growth); implemented scalable WebSocket streaming architecture, automated document ingestion and alert workflows, S3-integrated Excel updates, structured data normalization pipelines, and load-tested high-concurrency APIs using Locust for production-grade reliability and performance.

• Built a production-scale FastAPI + React AI annotation platform utilizing a multi-stage Gemini 2.5 Flash pipeline (promo filtering, confidence scoring, rule refinement, Dually verification) to automate truck image classification, reducing API costs by ~97% via prompt optimization and OpenCV mosaic batching; engineered a concurrent system with 10 parallel workers, Phoenix API key rotation, YODA rate limiting, checkpointing, streaming, 500-ad batch processing, integrated Selenium scraping, Trader API ingestion, PATCH-based production updates, and developed live dashboards with token-level cost tracking to process thousands of listings end-to-end with minimal manual intervention.

• Created a local mail server for the employees of DRDO named as drdo.local for their internal sharing of data and information.
• Used Hmailserver as the primary mail server for its implementation. Learned all the functionalities and technicalities of the HmailServer.
• Used Thunderbird app as email receiving and sending platform and used Roundcube with Xampp to host the mail server on the browser aside from thunderbird app.
• Learned about Virtual Machine for the testing purposes of the HmailServer.

• Developed a full-stack MERN website under the guidance of Mr. Anil Kumar, Senior Professor and Head of General Admin Office at DTU.
• The website features an Admin Portal, Employees Portal, and a user-friendly interface for students.
• Students can access important documents such as leave applications and employee information.
• Admin can assign tasks to employees, create new employee logins, change passwords and manage accounts.
• Employees are categorized by departments (e.g., Tender, DGHS, Basic Facilities), ensuring focused work assignments and have options to change their work assignments.
• Technologies used: HTML, CSS, JavaScript, React.js, Bootstrap, Node.js, Express.js, and MongoDB.