Deepak Kumar Soni

I don’t just build AI models; I architect intelligent systems that turn complex data into meaningful, real-world solutions. Currently, I am interning at the National Remote Sensing Centre (NRSC), ISRO, where I work at the intersection of artificial intelligence, satellite remote sensing, and geospatial analytics. My current flagship endeavour is CHATGS, an AI-powered geospatial intelligence platform integrating Google Earth Engine, satellite imagery, and conversational AI. It enables natural language-driven analysis of reservoir water spread, land cover, snow cover, and urban expansion using datasets including Sentinel, MODIS, and ISRO's EOS-04 SAR imagery. Beyond geospatial intelligence, I explore multimodal AI, computer vision, autonomous agents, and local-first AI systems, building voice assistants, real-time recognition applications, and modular hardware orchestration platforms using Python, FastAPI, React, and LLMs. My foundation in Medical Science and AI/ML, combined with diagnostic laboratory experience at Tata 1mg and mentoring students in Computer Science and Medical Laboratory Science, gives me a multidisciplinary perspective on engineering. My philosophy: Beyond the Black Box. I believe AI should be explainable, secure, scalable, and human-centered—not merely impressive in demonstrations, but dependable in practice. My core expertise spans LLM orchestration, RAG, computer vision, geospatial AI, remote sensing, data engineering, and full-stack development. I am also preparing for GATE DA to strengthen my mathematical and statistical foundations. My goal is simple: Build intelligent technology that doesn't just automate tasks, but understands context, bridges disciplines, and solves problems that genuinely matter.