End-to-end product ownership
Requirements, architecture, development, deployment, maintenance, iteration, and client communication.
Dinesh Bhattacharya
Software Engineer Agentic AI AI Systems
I build software that turns messy problems into working systems — increasingly with AI agents in the loop.
I build full-stack software, backend services, AI-powered applications, agents, and automation. My work sits at the intersection of practical product engineering and reliable AI systems.
About
I like building the parts that make software actually work.
I am a Software Engineer with hands-on experience across full-stack development, backend systems, REST APIs, AI integrations, and automation. I founded Vision Binary, worked directly with clients, and owned software from requirements and architecture through development, deployment, and iteration.
My current direction is Agentic AI and production-grade AI systems: applications that reason over context, use tools, connect with external systems, and automate meaningful workflows without losing engineering discipline.
What I Bring
Software engineering fundamentals with AI systems in view.
Full-stack and backend engineering
Interfaces, APIs, authentication, databases, async processing, caching, cloud deployments, and CI/CD.
AI-enabled automation
LLM applications, AI agents, RAG workflows, function calling, voice-based AI systems, and tool-driven workflows.
Experience
Building products, services, and AI-powered workflows.
Aug 2025 – Aug 2026
Founder & Software Engineer
Vision Binary - Kolkata, India
- Built and delivered full-stack software solutions for business clients, taking projects from requirements and architecture through development, deployment, and iteration.
- Designed and integrated AI agents, LLM-powered workflows, and voice-based AI calling systems to automate practical business processes.
- Developed backend services, REST APIs, third-party integrations, and automation workflows while owning implementation, deployment, and maintenance.
- Worked directly with 3 clients, translating business requirements into technical solutions and managing scoping, communication, and delivery alongside engineering.
Aug 2025 – Jul 2026
Software Engineer (Part-time)
Adyakshar Labs - Remote
- Designed, developed, and deployed full-stack applications and backend services using Node.js, TypeScript, Python, Express.js, Angular, PostgreSQL, MongoDB, Redis, Docker, and Linux.
- Integrated LLMs, AI Agents, OpenAI APIs, RAG, function calling, prompt engineering, and vector databases into applications and automation workflows.
- Developed REST APIs, JWT/OAuth2 authentication, asynchronous job processing, Redis caching, CI/CD workflows, and cloud deployments.
- Contributed across requirements, implementation, testing, debugging, deployment, and maintenance of production applications.
Selected Projects
Project work with clear technical substance.
KanbanFlow
Jira/Trello-inspired task management platform
- Problem
- Task workflows need clear movement, filtering, and persistent state without a heavy product surface.
- What I built
- A Kanban application supporting task creation, workflow management, filtering, and drag-and-drop interactions.
- Key detail
- Component-based Angular architecture with Signals, computed state, reusable components, client-side persistence, and Vitest tests.
PriceIQ
E-commerce price monitoring system
- Problem
- Historical e-commerce pricing data is hard to monitor without clear analytics and trend visibility.
- What I built
- An interactive price monitoring and analytics dashboard for analyzing historical pricing data.
- Key detail
- Implemented polynomial regression, OLS regression, moving averages, Z-score anomaly detection, and 95% prediction intervals without external ML libraries.
Detection of Phishing URLs
Nature-inspired phishing URL detection
- Problem
- Phishing URLs require fast, lightweight classification using URL-level signals.
- What I built
- A phishing URL detection system using Genetic Algorithms and Particle Swarm Optimization.
- Key detail
- Achieved 96.8% detection accuracy and designed the pipeline to process over 10,000 URLs per minute using lexical, structural, and domain-based features.
Building With AI
Where I am going deeper.
My current technical focus is Agentic AI: useful software where models can reason over context, use tools, call functions, interact with external systems, and support real workflows.
Hands-on experience
- LLM applications
- AI Agents
- OpenAI API
- RAG
- Function / tool calling
- Prompt engineering
- Vector databases
- AI automation workflows
Currently exploring
- Agentic AI architectures
- Reliable tool use
- Production AI systems
- Agent evaluation
- AI observability
- MCP
Technical Stack
A practical toolkit for shipping software.
Languages
TypeScript, JavaScript, Python, SQL, C/C++, HTML/CSS
Frameworks & Libraries
React, Angular, Node.js, Express.js, Flask, RxJS, Recharts, Scikit-learn
AI & ML
LLMs, AI Agents, OpenAI API, RAG, Function Calling, Prompt Engineering, Vector Databases, OLS Regression, Polynomial Regression
Databases
PostgreSQL, MongoDB, MySQL, Redis
Developer Tools
Git, Docker, Linux, Vite, Angular CLI, VS Code, Postman
Engineering
REST APIs, Microservices, CI/CD, JWT/OAuth2, System Design, DSA
Currently Building
Going deeper into Agentic AI.
I am currently focused on building reliable AI systems that can reason over context, use tools, interact with external systems, maintain state, and automate meaningful workflows.
Education
Computer Science foundation.
Indian Institute of Engineering Science and Technology (IIEST), Shibpur
Bachelor of Technology in Computer Science and Technology