Agentic AI Engineer
for New York Teams
LLM Applications · RAG Systems · Autonomous AI Workflows · FastAPI
I am Ramesh Kumar Das, a dedicated AI Full-Stack Engineer bringing over 6.5 years of experience in architecting and deploying resilient, production-ready web platforms and AI-driven SaaS applications. I partner with New York-based startups and established product teams to transform ambitious AI concepts into reliable software. My technical expertise spans advanced LLM integrations, sophisticated RAG pipelines, autonomous agentic workflows, FastAPI microservices, and highly responsive Next.js interfaces. Currently, my primary focus at Kliky AI is the development of WinstaAI—a comprehensive agentic SaaS platform that features intelligent multi-model routing, advanced media generation pipelines, workflow automation, and robust operational tooling.
The value I bring is practical engineering, not AI demos that fall apart in production. I focus on clean architecture, controlled API costs, secure data handling, reliable background jobs, useful admin controls, and deployment paths that make sense for real users.
AI and Full-Stack Engineering Stack
Technologies I use to plan, build, deploy, and maintain production AI applications.
AI Engineering Services for New York Companies
Hands-on AI development for SaaS teams, founders, agencies, and businesses adding automation to existing products.
🤖 Generative AI Product Development
I build AI SaaS products with LLM APIs, account systems, credits or billing logic, admin panels, usage analytics, and model orchestration. A typical build uses FastAPI or Django on the backend with a React or Next.js frontend for customers and internal teams.
🔍 RAG Pipeline Architecture
I design retrieval-augmented generation systems that help users query private documents, internal knowledge bases, product data, and support content. This includes document cleanup, chunking, embeddings, vector search, prompt design, access control, and answer quality checks.
🧠 Agentic AI Workflows
I create AI agents that can plan steps, call tools, use APIs, process files, and complete structured tasks. These workflows can support research assistants, internal operations, content pipelines, customer support processes, and multi-step business automation.
🎨 AI Media Generation Systems
I work with text-to-image, face swap, OCR, video AI, speech-to-text, and text-to-speech workflows using tools such as ComfyUI, Replicate, FAL.ai, and RunPod. I also help structure GPU usage so media generation features stay practical to operate.
🌐 AI SaaS Backend Development
I build backend systems for AI products, including queues, WebSocket streaming, rate limits, user permissions, usage metering, payment-aware credits, retries, logs, and background workers for long-running model tasks.
🔗 AI API Integration
I connect OpenAI, DeepSeek, Anthropic, Hugging Face, Replicate, and other AI providers to new or existing applications. My focus is clean integration, safe output handling, fallback logic, cost visibility, and prompts that can be tested instead of guessed.
Relevant Work Experience
Production experience across AI platforms, SaaS products, backend systems, and full-stack applications.
WinstaAI — Agentic AI SaaS Platform
I lead development across the AI product stack, including LLM workflows, media generation pipelines, backend services, dashboards, and admin tooling.
- Built agentic AI workflows for tool use, autonomous task execution, branching decisions, and multi-step orchestration with LangChain and custom backend logic.
- Integrated text-to-image, face swap, OCR, video AI, audio AI, and LLM routing across OpenAI, DeepSeek, Ollama, ComfyUI, and Replicate.
- Developed an AI operations dashboard for service control, user credits, billing visibility, usage monitoring, and admin management.
- Connected GPU inference services such as RunPod and FAL.ai with workflows designed for reliability and controlled operating cost.
Dibbly and Paperport — AI Video Translation and Content Generation
- Built multilingual AI video translation workflows using speech-to-text, text-to-speech, DeepL, and transcription processing.
- Implemented LLM-assisted content generation for converting unstructured inputs into presentations, summaries, and structured roadmaps.
- Worked on secure AI API communication, authentication, session handling, CSRF protection, and JWT-based access control.
Khantailor and PathHub
- Built and maintained full-stack platforms using React.js, Node.js, and Django.
- Added real-time features with WebSockets and Redis to support live updates across users.
- Improved backend performance through API optimization, caching, and database-aware engineering.
Frequently Asked Questions
Answers for New York founders, hiring managers, and product teams evaluating an AI engineer.
Need an Agentic AI Engineer for a New York Product?
I’m available for remote AI engineering roles, product builds, architecture consulting, and fixed-scope MVP work.