AI Engineer Intern @EZEU | CSE Graduate

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Bachelor of Engineering - BE, Computer Science
Bapuji Institute of Engineering & Technology, Davanagere
Science
Sir MV PU college
Join Vignesh and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal BossAI Engineer Intern @EZEU | CSE Graduate
CSE Graduate and Software Engineering Intern at ezeu, specializing in the intersection of high-concurrency backend systems, cloud deployment, and machine learning applications. Current Experience: AI Engineer Intern at ezeu, focusing on scalable product development and cloud architectures. AI & Document Intelligence (Ex-LogicBrackets): Architected seiton-ai-ocr-v1, integrating multi-modal LLMs (Qwen-VL) and remote PaddleOCR-VL via vLLM to hit a 92% text reconstruction accuracy rate. Engineered a production-grade RAG chatbot within SeitonDMS for highly accurate document querying. MLOps & Cloud Infrastructure: Skilled in multi-stage Docker containerization and deploying high-throughput, low-latency API architectures directly onto Cloud GPUs. Core Tech Stack: Python, FastAPI, PyTorch, Multi-Modal LLMs, vLLM, Docker, Cloud GPUs, Vector Databases, OpenCV, Git.
Bengaluru
AI Engineer Intern at EZEU (OPC) INDIA PVT LIMITED, Jun 2026 - Present · 2 mos, Bengaluru, Karnataka, India
Software developer Intern at LogicBrackets, Jan 2026 - Jun 2026 · 6 mos, Bengaluru, Karnataka, India, Scalable Architecture: Designed an object-oriented, production-ready FastAPI Service-Router backend supporting asynchronous, high-concurrency multi-page document processing. SeitonDMS RAG Chatbot: Engineered an advanced Retrieval-Augmented Generation (RAG) chatbot within SeitonDMS, enabling high-accuracy contextual information querying directly from drive-uploaded documents. Document Intelligence: Integrated PP-Structure (DSR) alongside local Qwen-VL multi-modal models for zero-shot entity recognition and context-aware Key-Value extraction, achieving a 92% layout reconstruction accuracy rate. Optimization & Deployment: Containerized backend services utilizing multi-stage Docker builds and deployed the infrastructure onto Cloud GPUs, leveraging the vLLM framework and custom LRU caching to eliminate CUDA memory leaks.
Bachelor of Engineering - BE, Computer Science, Bapuji Institute of Engineering & Technology, Davanagere
Science, Sir MV PU college