Initializing...
Hello, I'm
Full Stack AI Developer crafting intelligent systems — from multimodal OS agents to agentic career platforms. Based in Pune, India.
Hello! I'm Siddhesh Asati, a passionate Full Stack AI Developer with a strong interest in building meaningful intelligent systems. I specialize in transforming complex ideas into elegant, production-ready AI applications.
With a foundation in machine learning, generative AI, RAG pipelines, and full-stack web development, I've built projects ranging from multimodal OS-level automation agents to AI-powered career platforms.
When I'm not engineering code, you'll find me directing short films or out on a long drive — creativity fuels both passions.
CGPA: 7.91/10 · Data Science as Honor subject. Focused on advanced computing, algorithmic problem-solving, and data-driven decision-making.
Percent: 91.5 · Maharashtra State Board · Simultaneous JEE preparation at ICAD Institute.
Percent: 81.67 · CBSE · Information Technology as Honor subject.
Building end-to-end RAG pipelines and scalable LLM applications with a React.js frontend. Acquiring industry expertise in model fine-tuning and building robust AI solutions.
Engineered an Automated Compressor Performance Prediction system using regression models to forecast efficiency with 94% accuracy. Built a React.js analytics dashboard with backend logic for real-time analytics.
Hands-on experience in AI, Machine Learning, and Deep Learning using Python. Developed neural network applications and data-driven pipelines.
Agentic AI LLM System
Multimodal AI assistant designed as Agentic AI System and real-time web search. Integrates LLMs with system-level APIs for seamless task execution — open, close apps, generate content, search the web, integrated RAG pipelines.
AI Career Platform
AI-powered career platform with real-time job listings, automated resume matching using vector similarity, and interactive AI interview simulation powered by LangChain and LLMs.
AI-Driven Banking Application
A secure, modern web-based banking platform integrated with AI-driven features including predictive spending analytics, an intelligent chatbot for customer queries, and automated anomaly detection for transaction security.
Agentic OS Automation
Advanced Agentic OS automation tool built for seamless, script-driven workflow executions. Enables natural language-to-system-action pipelines leveraging LangGraph agentic loops.
High-performance FastAPI system leveraging a two-stage retrieval and reranking pipeline with BGE embeddings and Qdrant Vector DB for semantic financial querying. Features robust SQLite-backed Role-Based Access Control (RBAC), JWT authentication, and automated PyMuPDF/Pandas parsing pipelines.
Content-based recommender system using cosine similarity on movie metadata — genre, cast, plot. Built with Scikit-Learn and deployed via Streamlit.
A deep dive into AI-driven chatbot systems — architecture, NLP pipelines, and deployment strategies for simulating human-like conversation.
Read More →How recommender systems power personalized experiences — from Netflix suggestions to Amazon product recommendations, with code walkthroughs.
Read More →Real-time passenger tracking and dynamic seat allocation using QR code technology — a full system design breakdown.
Read More →Simplifies ticketing and passenger administration via QR code technology. Customers booking through a web platform receive digital QR tickets validated at coach entrances by external scanners.
Read PaperRemote sensing ship recognition for ocean security monitoring addressing challenges of multi-scale targets, arbitrary direction, dense arrangement, and sea fog interference.
Read PaperOpen to Full-time roles, collaborations, and interesting AI/ML projects. Let's build something great together.