profile-pic

Dipak Kumar Tomar

Experienced Full-Stack Developer with expertise in frontend, backend, and mobile development using React Native. Skilled in creating seamless user experiences and scalable backend systems, delivering comprehensive end-to-end solutions. Passionate about leveraging the latest tools and staying ahead of industry trends to build impactful, innovative applications.
  • Role

    AI/ML Engineer - LLM & Agentic Systems

  • Years of Experience

    3.8 years

  • Professional Portfolio

    View here

Skillsets

  • NLP
  • Redis
  • rag
  • Python
  • Prompt Engineering
  • PostgreSQL
  • pipelines
  • OCR
  • OAuth
  • Redux toolkit
  • Next.js
  • Neo4j
  • Microservices
  • MCP
  • LLM
  • LangGraph
  • LangChain
  • Supabase
  • Zustand
  • Zod
  • Websockets
  • Vercel
  • vector search
  • Typesense
  • Tool calling
  • Tailwind CSS
  • Lambda
  • Structured output
  • SSE
  • SQLite
  • Serverless
  • Sentry
  • S3
  • REST
  • React.js
  • Cloudfront
  • CI/CD
  • AutoGen
  • asyncio
  • API Gateway
  • Agentic AI
  • Agentic AI
  • TypeScript
  • crewAI
  • Node.js
  • MongoDB
  • Firebase
  • AWS - 3 Years
  • React Native
  • JavaScript - 3 Years
  • React Native
  • JavaScript
  • Data Modeling
  • Docker
  • EC2
  • embeddings
  • evaluation workflows
  • Expo
  • Express.js
  • FastAPI
  • Firestore
  • Flask
  • GitHub Actions
  • GraphQL
  • GuardRails
  • JSON
  • JWT

Professional Summary

3.8Years
  • Feb, 2025 - Present1 yr 9 months

    AI/ML Engineer - LLM & Agentic Systems

    Techype
  • Feb, 2023 - Jan, 20251 yr 11 months

    Software Engineer - Mobile & Data Products

    Campus 365
  • Aug, 2022 - Feb, 2023 6 months

    Software Engineer - Platform & Mobile Applications

    Till POS Technologies

Applications & Tools Known

  • icon-tool

    React Native

  • icon-tool

    Redux

  • icon-tool

    Loopback

  • icon-tool

    AWS Lambda

  • icon-tool

    AWS S3

  • icon-tool

    Monorepo

  • icon-tool

    ReactJS

  • icon-tool

    Next.js

  • icon-tool

    TypeScript

  • icon-tool

    Redux

  • icon-tool

    Firebase

  • icon-tool

    React

  • icon-tool

    Redux-Saga

  • icon-tool

    AWS

  • icon-tool

    REST APIs

  • icon-tool

    React Navigation

  • icon-tool

    React Query

Work History

3.8Years

AI/ML Engineer - LLM & Agentic Systems

Techype
Feb, 2025 - Present1 yr 9 months
    Designed and shipped multi-agent LLM workflows using LangGraph, AutoGen, CrewAI, and Python for adaptive learning, document automation, research, and operations orchestration. Built an MCP server as the central tool-orchestration layer with JWT-authenticated SSE transport, role-scoped tool registry, structured JSON-RPC interactions, and Zod schema validation. Architected an agentic chat system with dynamic intent routing across 40+ intent types, using classifier-based state routing to select specialized workflows for search, outreach, pipeline management, and research. Implemented memory-aware AI workflows using -style retrieval patterns, structured outputs, guardrails, and validation checks to make agent responses safer and more reliable.

Software Engineer - Mobile & Data Products

Campus 365
Feb, 2023 - Jan, 20251 yr 11 months
    Built production React Native/TypeScript mobile features for education workflows with real-time Firestore data models for attendance, grades, and student operations. Integrated Firebase Cloud Messaging and real-time sync patterns to support event-driven notifications, live data updates, and reliable mobile state management. Collaborated with product/backend teams on API contracts, data validation, release stability, and performance improvements across mobile modules used by schools and institutions. Mentored 2 junior developers, led code reviews, and established reusable component/state-management practices for maintainable production code.

Software Engineer - Platform & Mobile Applications

Till POS Technologies
Aug, 2022 - Feb, 2023 6 months
    Built cross-platform iOS/Android payroll workflows for Payroo, an Australian payroll and payslip platform covering tax, superannuation, single-touch payroll, and employee onboarding. Developed React Native, TypeScript, and Redux features for employer/employee workflows involving sensitive financial data, form validation, and API integration. Strengthened production engineering foundations relevant to AI product delivery: reliable data flows, secure user workflows, asynchronous API handling, and mobile release quality.

Major Projects

4Projects

Maple Drive - AI Agent Platform with MCP & Personalized Memory

    Built MCP server for tool orchestration and memory retrieval with JWT-authenticated SSE, JSON-RPC, role-scoped registry of 30+ tools, and standardized agent-tool contracts. Architected Lucy AI agent with dynamic intent routing across 40+ intent types, chaining specialized workflows with 5-10+ tool calls for recruiting, outreach, search, and research tasks.

AI Medical Quotation Processor

    Built a multi-agent OCR + LangChain pipeline to extract, classify, and validate line items from unstructured medical quotation documents, reducing manual entry for insurance/billing teams. Deployed serverless AI processing on AWS with structured-output validation against business rules before downstream handoff.

AI Insurance Verification System

    Automated insurance policy verification using OCR/NLP workflows to extract policy fields, cross-check client rules, and flag discrepancies for operations review. Integrated verification service with client systems through REST APIs; containerized with Docker and deployed on AWS for repeatable production delivery.

Roomi - Marketplace Search/Data Platform

    Led migration of a live 1 lakh+ user marketplace from Bubble.io to a Next.js + Expo monorepo, mapping legacy data into normalized Supabase schemas for web, iOS, and Android. Implemented production-grade search/filtering, payments, identity flows, media storage, observability, and scalable data workflows useful for -heavy AI products.

Education

  • Bachelor of Technology (B.Tech), Mechanical Engineering

    Punjab Technical University (2022)

A Few Screening Questions Before You Begin

Yeah. Hi. So my name is Kumar Thomar, and I have done my graduation from Chandigarh Engineering College in 2022. And during my graduation, in 2020, I joined a company called I Leave It Good. And that company basically works on services, providing services in IT services, freelancing, by bidding on sites like Upwork and freelancer.com. Then I had to work on multiple projects, including ecommerce and education apps, which I have done in React Native. After that, I joined a company called Till Post Technologies. I worked as a software engineer, working on multiple projects. It's a product-based company, with only one project, called Peru, which is a chat solution. And after that, I joined a company called Campus Health, it's also a product-based company, an ERP solution company.

Device location service test its function tools, accuracy and efficiency. Okay. So I haven't used a package called React Native to access location services. So I don't have much experience with it. I just want to say that we can move on. I won't use it much. According to the documentation, I have done that.

Seamless launch, we can check if there are some errors or not in the build or cradle or in iOS, like in, to export. We can check that, if it's running properly. So that we can do.

Offline capability and data synchronization. Yeah. So through for offline capability, we can use React Native Offline library. Otherwise, we can use AsyncStorage to store the data in the cache. And then if the Internet is not working, then it will show the data from the cache. That's how we can handle offline capability and data synchronization.

And scale, straight into large meetings. So we can use, basically, Redux to manage and scale state in large React Native applications. Redux basically provides, like, you can create slices in your Redux toolkit to manage the state.

For push notification, you can do this through Firebase. You can create a project and then install the library for the active voice notification through that, and implement it. For both iOS and Android, there are different steps required.

You come across a word and you're interested. You can use forward expressing potential and state updates. So, you can remove the second one. If you want to, like, update it multiple, like 2 times. You can do previous, then set it down, bracket previous and then previous plus 2. So it will make it increase by 2 times because basically it will just synchronously update the state. So there is a zero value in the next state, and then will be 1 and 1 again, 2 times. That's not how it works. Yet, it's native or yet native. Both are the same.

I don't know about this. Like, I know it's saying irritants. So, the doc extends an animal. So, it will. I don't have to provide this.

Application with existing native modules. So native model is a process. Like, we can write code, and then we can export that as our native model, and then you can use a native model from a library called React Native.

We can use the electronic UI next. Yeah. We can create multiple files for English, json, en.json, and then light end dot JSON, highlight dot JSON. These things we can do.