Hi, I'm Snehanshu
Machine Learning, Backend, RAG & Agentic AI Specialist
Crafting digital experiences with cutting-edge Backend Architecture, RAG, Agentic AI, LLMs, and seamless Full-Stack Solutions.
Crafting digital experiences with cutting-edge Backend Architecture, RAG, Agentic AI, LLMs, and seamless Full-Stack Solutions.
I'm a passionate Software Engineer with deep expertise in Backend Development, Machine Learning, Computer Vision, Systems Design, Agentic AI, RAG, and DevOps, focused on building and delivering seamless user experiences.
Projects Experience
Work Years Experience
GPA: 3.86
GPA: 3.89
An AI-powered visual assistance platform that empowers everyone through real-time scene understanding, voice interaction, and intelligent image capture.
A powerful application that allows users to upload PDFs, URLs, and Audio files, then chat with an AI assistant that can search through and reference the uploaded content.
A large-scale TikTok data analysis pipeline using NLP and vector databases to uncover sentiment trends, user behaviors, and community insights.
Browser-Vault turns your entire web browsing history into a fully searchable memory. Instead of just saving URLs, it captures full page content—text, images, URLs, titles, and timestamps; creating a rich, AI-powered archive of everything you've seen online.
An AI-Agent that instantly locates specific topics across multiple PDFs. It understands natural language queries and directly opens the exact page in the right document. This tool streamlines research by combining semantic search with an interactive, navigable PDF viewer.
Benchmarked Trie, Binary, and Linear search algorithms on a 10M-record dataset to uncover real-world performance beyond theoretical complexity.
Chrome extension that captures screenshots and uses Google Gemini AI to analyze and explain anything visible on screen. From text explanations to image analysis, code debugging to visual Q&A - get instant AI-powered insights for any screen content.
An interactive N Queens game that lets you choose your board size and challenge yourself to place N queens safely without any attacking each other. A fun and visual way to explore this classic combinatorial problem in chess and algorithms.
A simple yet powerful tool to take control of GIFs: pause, play, and seek any frame! Available as both a web app and Chrome extension. The web version supports uploads and direct URLs, while the Chrome extension auto-detects GIFs on any page and overlays playback controls.
Mentoring 6 teams in designing and delivering real-world software systems, driving requirements clarification, technical architecture, and stakeholder-aligned execution.
Headed and built a state-of-the-art RAG-based “PDF Explorer” enabling expert-level, multi-modal document search, setting a new standard for AI-driven enterprise retrieval, with successful onboarding of India’s largest fashion brand.
• Designed and implemented a high-performance, multi-stage Computer Vision classification system for e-commerce catalog processing, classifying fine-grained attributes (e.g., color, sleeve length, fabric) for fashion items.
• Architected and built a fault-tolerant Dropbox image uploader service, improving throughput 4x by addressing scaling bottlenecks.
• Led development of a high-throughput indexing pipeline integrating Qdrant and Typesense for recommendations system.
• Optimized legacy codebases, improved documentation, debugged, and enhanced various workflows to boost overall performance.
Developed a custom Python DMN rule engine with rule manipulation, integrated DynamoDB-based versioning for rollback, and enabled DMN file export and reverse engineering for seamless data recovery.
Mastered Docker to create production-ready Business-Central images with RBAC, built lifecycle automation using Kie-Server REST APIs, and optimized Python request handling for AWS Lambda with clear client documentation.
• Data Tech Team: Enhanced autosuggest search with category classifiers and probabilistic N-gram models trained on 100K+ phrases, deploying scalable APIs and scripts.
• Visual Tech Team: Improved dress category prediction using advanced feature extraction and developed scalable image processing pipelines for bulk operations.
I'm always excited to collaborate on innovative projects and explore new opportunities. Feel free to reach out through any of the channels below!
srj.snehanshu@gmail.com
Los Angeles, CA