Hello, I'm
I build AI agents, integrations, and agentic AI systems.
Every system I deliver ships with automated tests, structured logging, and real-time monitoring built in. You get a Project Status Dashboard from day one — check test results, view logs, monitor health, and track API costs anytime during development.
View My WorkA little bit about who I am and what I do.
I'm a Software & AI Developer from India with a One Year Post Graduate Diploma in Artificial Intelligence & Machine Learning. My last full-time role was onsite in Munich, Germany, working for a reputed automotive company where I contributed to autonomous car development.
I specialize in building AI agents, agentic AI systems, and AI integrations that solve real business problems — from RAG-powered document assistants and conversational lead bots to multi-agent task runners and AI-driven CRMs.
Below you'll find live, interactive demos of the AI-powered applications I've built. Each one is a fully functional, production-ready system you can explore.
Technologies and tools I work with.
Live demos you can interact with. Click a card to open the app.
RAG-powered document assistant that lets you upload PDFs, Word docs, and text files, then ask natural-language questions and get accurate, source-cited answers from your documents.
Multi-channel AI lead qualification bot with a real-time dashboard. Automatically engages visitors, qualifies leads with intelligent conversation, and hands off to human reps with full context.
Upload spreadsheets and data reports, ask questions about your numbers in plain English, and get AI-generated charts and insights with auto-selected visualizations.
Full-stack CAE surrogate platform covering the complete ML engineering workflow: parametric FEA data generation → physics-informed neural network training → ONNX export → REST API serving → inverse design optimisation. Given a stress or displacement constraint, the app finds the optimal geometry and material combination across steel, aluminium, titanium, and copper. Surrogate predictions run at <0.2 ms vs minutes for FEA, achieving R²=0.997 with 500 training samples.
Let's work together. Reach out through any channel below.
I'm open to freelance opportunities, collaborations, and interesting projects. Don't hesitate to get in touch!