“Building cool stuff, one idea at a time.”
I'm Rohan Sinha, an AI/ML developer from Bangalore, India, with an engineering background and a strong passion for building intelligent systems. My journey revolves around deep learning, large language models, and turning ideas into working AI products.
I enjoy exploring how machines learn, reason, and solve real-world problems. Through DeepNova, my AI-focused initiative, I work on making complex machine learning concepts easier to understand and building practical AI solutions. I'm also proud of developing Coding_Model, a lightweight language model designed to explore the possibilities of locally running AI.
From experimenting with compact language models to building AI projects such as DeepNova and a chest X-ray analysis system, I focus on learning by building, testing, and improving. My goal is to create efficient, accessible AI systems that make a meaningful impact, one experiment at a time.
DeepNova AI is my vision to make artificial intelligence and deep learning accessible to everyone through intuitive learning experiences, innovative AI projects, and real-world solutions.
Currently mastering deep learning, LLMs, and agentic AI to build intelligent systems that solve real-world problems.
Learning to deploy production level AI systems.
My skills, honestly ranked — from the technologies I use to build AI projects like DeepNova and my coding model, to the tools I’m mastering and the concepts I’m exploring next.
Engineering robust machine learning pipelines, validating computer vision datasets, and solving real-world AI challenges in production.
Focusing on AI training, data quality assurance, and dataset validation for large-scale computer vision and machine learning models.
Engineered predictive classification models and end-to-end machine learning data workflows under the joint Microsoft and SAP CSR initiative.
Real-time telemetry pulled directly from the GitHub API — tracking active repositories, open-source AI models, algorithmic problem solving, and ongoing engineering commits.
Live API sync active · Showing top repositories from github.com/Rohansinha2003
51M parameter lightweight language model trained for code completion and local inference, published on Hugging Face.
AI platform making artificial intelligence and deep learning accessible through intuitive learning experiences.
Deep learning computer vision system for pulmonary disease detection from chest radiographs.
Autonomous AI agent framework with dynamic tool calling and multi-step reasoning.
Retrieval-augmented generation pipeline with vector search and document grounding.
Legal tech AI workflow and document processing solution.
Personal engineering portfolio showcasing AI/ML projects and interactive 3D visualizations.
Click any card for details, or open the full case study for the problem, constraints and trade-offs.
An AI-focused platform making artificial intelligence and deep learning accessible through intuitive learning experiences and real-world tools.
In a hurry? Skip the scroll and ask the assistant. It only knows the resume and projects.
An algorithm to crack, a product to ship, or the latest thing you are into. Drop a line.