ComputerUSE
AI computer control agent
An autonomous agent that combines screen understanding with LLM reasoning to plan and execute UI actions.
Vision models, LLM planning, UI automation
View GitHub→
Deepesh Gupta
AI/ML Engineer
Deepesh
Gupta
Building practical AI systems across LLM applications, retrieval workflows, and computer vision.
Currently focused on object detection, active on Kaggle, and happiest building with coffee nearby and music in the background.
Currently
Role
AI Research Engineer at Magure India Pvt. Ltd.
Focus
Object detection, practical AI systems, and applied ML workflows
Outside work
Kaggle experiments, coffee, and music
Background
B.Tech in Computer Science (AI & ML), KK Modi University
Availability
Open to AI/ML engineering and applied LLM roles
About
AI/ML engineer specializing in computer vision and generative AI.
I've worked across object detection, RAG systems, document intelligence, and research-oriented AI tooling. My focus is building practical systems that combine strong experimentation with production-minded engineering.
Experience
AI Research Engineer Intern
Magure India Pvt. Ltd. · Feb 2026 – Present
Working on applied AI research and intelligent systems with a focus on practical experimentation and product-oriented development.
Localization Engineering Intern
Andovar · Jul 2024 to May 2025
Automated file processing workflows across JSON, XML, TMX, PPTX, and DOCX, reducing manual work by 40% while improving delivery speed and consistency in localization pipelines.
Projects
ResearchDigest
AI research analysis platform
Natural-language search and AI summarization for research papers, with semantic retrieval and automated insight extraction.
Hugging Face, semantic search, full-stack web architecture
View GitHub→QuickTrack
Real-time object detection
A Streamlit application for YOLOv8-based object detection and tracking with interactive controls and exportable results.
YOLOv8, Streamlit, real-time inference
View GitHub→MultiPDF ChatBot
Document intelligence system
A LangChain-based RAG workflow for question answering across multiple PDFs using FAISS and semantic chunking.
LangChain, FAISS, PDF processing, retrieval pipelines
View GitHub→Capabilities
LLM & Retrieval
RAG systems, LLM fine-tuning, prompt engineering, LangChain, LlamaIndex, Hugging Face, FAISS, Pinecone, and Chroma.
Computer Vision
OpenCV, YOLO, object detection, tracking pipelines, semantic chunking, and document intelligence workflows.
ML Engineering & Cloud
Python, SQL, FastAPI, Docker, Kubernetes, MLflow, GCP, Vertex AI, and Cloud Run deployment workflows.
Data & Experimentation
TensorFlow / Keras, Scikit-learn, XGBoost, Pandas, NumPy, Matplotlib, Seaborn, and Plotly.