I'm Vinaya Kumar, a Machine Learning Engineer and AI-focused Software Engineer passionate about building intelligent, scalable, and production-ready software.

My work combines Machine Learning, Generative AI, and Full Stack Development to solve real-world problems. I enjoy taking products from an idea to deployment by designing clean architectures, building efficient APIs, integrating LLMs, and creating intuitive user experiences.

• Designing end-to-end AI applications with modern frameworks.
• Fine-tuning and deploying LLMs (Mistral, Qwen, Hugging Face).
• Building backend services using FastAPI, Python, and REST APIs.
• Developing clean interfaces with React and Tailwind CSS.
• Implementing RAG pipelines for contextual search & query handling.
• Optimizing AI inference pipelines using tools like GGUF & QLoRA.

Engineering Creed: Write maintainable code. Build scalable systems. Solve meaningful problems.

• Machine Learning ⭐⭐⭐⭐⭐ Scikit-learn • TensorFlow • NumPy • Pandas End-to-end ML model development, feature engineering, preprocessing, and validation.
• Deep Learning ⭐⭐⭐⭐☆ TensorFlow • CNNs • Keras Built CNN-based facial emotion recognition models with high classification accuracy.
• Natural Language Processing ⭐⭐⭐⭐⭐ Hugging Face • DistilBERT • Transformers Customer sentiment analysis, text classification, and inference pipelines.
• Generative AI & LLMs ⭐⭐⭐⭐⭐ Mistral • Qwen • Prompt Engineering Built AI assistants powered by locally hosted LLMs, fine-tuning, and inference optimization.
• Retrieval-Augmented Generation (RAG) ⭐⭐⭐⭐☆ LangChain • Vector Search • Embeddings Retrieval-Augmented Generation pipelines and contextual AI systems.
• Model Fine-Tuning ⭐⭐⭐⭐⭐ QLoRA • PEFT • PyTorch Fine-tuned Qwen2.5-3B on custom domain datasets.
• Computer Vision ⭐⭐⭐⭐☆ OpenCV • CNNs • Facial Landmarks Facial emotion recognition using image processing and deep learning.
• Model Deployment ⭐⭐⭐⭐☆ FastAPI • Docker • REST APIs Production-ready AI inference services with low latency.

I believe great software is more than writing code. My development approach focuses on:

• 🧩 Problem Discovery: Understanding business problems before writing code.
• 🏗️ Architectural Scale: Designing modular, scalable, and extendable systems.
• 🧹 Clean Code: Adhering to SOLID and clean-code principles.
• 👥 User-Centricity: Building seamless, intuitive, and responsive interfaces.
• ⚡ High Reliability: Optimizing latency, resource usage, and reliability.

• 🤖 AI Engineering: Machine Learning, Generative AI, LLMs, NLP, Computer Vision
• 💻 Software Engineering: Backend Development, Full Stack Development, API Engineering
• 🤝 Collaboration: Open Source, Research Projects, Hackathons, Startup Teams

• 🤖 Agentic AI Systems & Multi-Agent Architectures

• 🔍 Advanced RAG Pipelines & LLM Evaluation Frameworks

• 🔌 Model Context Protocol (MCP)

• ☸️ Kubernetes for AI Workloads & Infrastructure