Praveen Kumar S.

I build AI-enabled products, knowledge assistants, and agent workflows, from the web or mobile interface to the backend, integrations, and deployment. I can own a complete build or a defined part of your existing product. I'm Praveen, a software engineer with 7+ years of experience and hands-on CTO experience. I'm Top Rated Plus on Upwork with 100% Job Success. My work spans healthcare AI, audio products, fintech, and commerce, with a strong focus on connecting AI capabilities to applications people can actually use. What you can hire me to build • AI products and features: a new application, an assistant inside your existing software, or a workflow that connects your data, APIs, and business logic. • Knowledge assistants and agents: retrieval-grounded answers, tool routing, conversation state, and fallback behavior, connected to a usable interface. • Real-time chat and voice: streaming conversations and audio experiences across web, mobile, and backend services. • The supporting product: React or Next.js interfaces, React Native apps, authentication, backend APIs, asynchronous workers, and deployment. • Improvements to an existing system: focused work on unreliable sessions, failed background jobs, slow data access, integrations, and release pipelines. Relevant work As CTO of a confidential healthcare AI platform, I lead a three-person engineering team while writing code across the backend, AI workflows, and applications. I built LangGraph workflows for symptom intake, retrieval, reasoning, tool routing, safety fallbacks, and response composition. The retrieval work brought together structured health records, clinical notes, and medical guidelines through embeddings and vector search. On that same platform, I developed real-time chat and voice services with Node.js and WebSockets, integrated virtual-assistant experiences using Unreal MetaHuman and Pixel Streaming, and owned React Native architecture across mobile, Windows, and Android kiosks. The implementation details matter here. I've worked on session completion that waits for persistence and final audio delivery, notification workers with retries and graceful shutdown, and regression checks for clinical guidance across translations. Those are the kinds of issues I look for when a feature needs to behave consistently beyond the happy path. At Apollo 24/7, I built in-app VoIP communication between patients and support teams and reworked deep linking, referral flows, and attribution integrations. At Audvice, I built mobile and web capabilities from zero to one, including audio capture, content workflows, and AWS Lambda processing for transcription and media optimization. At Vested Finance, I modernized React Native architecture, rebuilt portfolio dashboard experiences in Next.js, and introduced Redis caching for frequently requested financial data. How I work I start with the user journey, the current system, and a clear definition of done. Then I break the work into reviewable milestones, explain trade-offs, and stay involved through implementation, testing, and deployment. You can bring me a product to build or a specific workstream to own alongside your team. My core stack is TypeScript, Node.js, React, Next.js, React Native, and PostgreSQL. For AI workflows, I work with LangGraph, LlamaIndex, retrieval-augmented generation (RAG), and Azure OpenAI. I use queues, caching, Docker, and AWS where they fit the product's needs, with attention to authentication, retries, timeouts, and useful operational logs. Message me on Upwork with what you're building, where the project stands, and the main challenge you want to solve.