I'm Samin Saikia — a passionate full-stack developer and AI enthusiast based in Assam, India. I love building digital products that solve real problems and push the boundaries of what's possible with modern technology.
My journey spans across web development, machine learning, and agentic AI systems. Whether it's crafting pixel-perfect UIs or training deep learning models, I bring the same level of dedication and attention to detail to every project I take on.
End-to-end web solutions that captivate users. From responsive landing pages to complex web apps built with modern frameworks and best practices.
Custom software solutions architected for performance and maintainability. Clean, well-tested code that stands the test of time.
Seamless integration of frontend beauty and backend power. REST APIs, databases, authentication — the full package delivered with precision.
Predictive models and ML pipelines that transform raw data into actionable insights. Classification, regression, clustering, and more.
CNNs, transformers, LLMs — building and fine-tuning deep learning architectures for computer vision, NLP, and beyond.
Designing and deploying agentic AI workflows with tool-use, planning, and multi-step reasoning. The next generation of intelligent automation.
Turning raw datasets into clear narratives through statistical analysis, visualization, and exploratory data science using Python and modern libraries.
Agentic AI
MAXUS is a modular AI assistant built using the Groq API, combining large language models, multimodal vision processing, voice interaction, and persistent memory into a unified AI console. Inspired by cinematic AI systems, This is not just a chatbot. This is a self-contained AI engine.
Agentic AI
Architect → Build → Test → Document. A fully automated, human-in-the-loop AI pipeline that takes a project idea from concept to documented codebase using specialized LLM agents
Agentic AI
A locally-hosted AI research agent that autonomously browses, reads, and synthesizes information into comprehensive reports — powered by Groq/Compound.
Deep Learning
🧠 Brain MRI classifier built with MobileNetV2 transfer learning — 88.1% accuracy, TFLite optimized, deployed on Render. Full pipeline: Google Colab training → Flask API → live web app. No local environment used.
Machine learning
ML-powered product review sentiment analyzer — TF-IDF + Linear Regression trained on 190K+ Flipkart reviews, served via Flask REST API with a stunning frontend. Live on Render.
Web app
GitHub analytics dashboard built with Flask, SQLAlchemy and Chart.js. OAuth login, background sync, commit heatmap, language distribution and hourly activity charts.
A curated showcase of my best work, skills, and professional achievements.
I'm always open to discussing new projects, creative ideas, or opportunities to be part of something amazing. Drop a message and let's build something great together.