Vineet Kumar
about

I'm a Computer Science and Engineering student pursuing my bachelor's at IET Lucknow with a strong interest in AI, software development, and problem-solving. I enjoy exploring Generative AI, Web Development, Machine Learning, and API development while continuously learning and refining my skills. Currently, I'm gaining hands-on experience as a Data Science Intern at Tensor Solution while freelancing.

experience
2025
Data Scientist Intern Tensor Solution Tech
• As a Data Scientist Intern at Tensor Solution Tech Pvt. Ltd., I am contributing to AI-driven search solutions. I explore and implement AI techniques, including agents and tools to enhance search capabilities.
• Developed advanced search systems supporting vector, keyword, image-to-image, text-to-image, speech-to-speech, speech-to-text, and text-to-speech queries, enabling fast and multi-modal retrieval.
• Designed and deployed robust FastAPI-based microservices for tasks like image captioning, OCR, translation, language detection, and audio-text conversions.
• Integrated LLMs and ML models from Hugging Face and other sources to power intelligent tool-calling and agent-style reasoning in conversational search pipelines.
• Built a dynamic agent-based search framework that gathers, analyzes, and synthesizes information before responding — mimicking human-like query reasoning.
• Created and integrated a Shopify plugin to bring these AI-powered services directly into merchant workflows for enhanced product discovery.
Technologies - Python, FastAPI, MongoDB, Typesense, LangChain, Transformers, HuggingFace, Docker, Pydantic
projects
Ask PDF
A FastAPI-based application that uses LangChain to query PDFs via RAG while maintaining conversational context. It features agentic tools like Web Search and Math and supports multilingual speech input with Whisper AI for transcription and Mistral-7B as a Language Model.
Tech Stack: LangChain, FastAPI, HuggingFace, LangChain Tools and Agents, Javascript
Searching Simulation Using Typesense
A streamlit application to demonstrate searching from desired large dataset (xlsx, csv, json, jsonl) using typesense. Offers voice search using Whisper and Typed search query.
Tech Stack: Typesense, Streamlit, Python
Club Events Registrations & Management
A platform for college-wide club event organization, handling both individual and team- based event registrations, with admin controls for managing events and registrations data.
Tech Stack: Django, REST API, JavaScript, HTML/CSS, SQLite