Full-Stack & AI Engineer (LLMs, RAG, Agents)

Satendra Kumar

Full-Stack & AI Engineer Python TypeScript React LLMs · RAG · Agents · MCP 5 Years of Experience Gurugram, India — Open to relocation: Bengaluru · Hyderabad · Pune H +91-700******* # ********.**********@*****.*** ï linkedin.com/in/satendrakumar8 § github.com/Satendra8 SUMMARY

Full-stack engineer with 5 years of experience building secure, scalable B2B SaaS and real-time systems in Python, FastAPI, Node.js, TypeScript and React. Builds LLM-powered applications: agentic workflows with tool call- ing, RAG pipelines (chunking, embeddings, vector retrieval with pgvector, grounded answers) and streaming chat UIs. Completed Scrimba’s AI Engineer Path (agents, RAG, MCP, context engineering, evals, guardrails). Strong in system design, API architecture, authentication and access control, PostgreSQL performance and event-driven sys- tems. Owns the full lifecycle from requirements and design docs through deployment and production support, and has led and mentored teams of 5–6 engineers.

TECHNICAL SKILLS

Generative AI: LLM Applications, Agentic Workflows, AI Agents, Tool / Function Calling, Model Context Protocol

(MCP), RAG (Ingestion, Chunking, Embeddings, Retrieval, Grounding), Vector Databases (pgvector), Prompt & Con- text Engineering, Streaming Responses, Guardrails, LLM Evaluation (Evals), LangChain, LangGraph, Vercel AI SDK LLM Providers: OpenAI, Anthropic Claude, Google Gemini, OpenRouter Languages: Python, TypeScript, JavaScript, SQL

Backend & APIs: FastAPI, Pydantic, Node.js, NestJS, Express.js, REST APIs, Microservices, WebSockets, Socket.IO Frontend: React, TypeScript, HTML, CSS, Tailwind CSS, Vite Databases & Caching: PostgreSQL, Supabase, MySQL, MongoDB, Redis, TypeORM, Database Design, Query Optimisation, Read Replicas

Distributed Systems: RabbitMQ, AWS SQS / SNS, Asynchronous Processing, Background Workers, Event-Driven Architecture

Security: Authentication & Authorization (JWT, Sessions), Role-Based Access, Rate Limiting, Tenant Data Isolation, Data Privacy

Cloud, DevOps & Observability: AWS (Lambda, ECS, EC2, S3), Docker, Git, GitHub Actions, CI/CD, Prometheus, Grafana, Loki

Engineering Practices: System Design, Application Architecture, HLD, LLD, SOLID, Design Patterns, Automated Testing (Jest), TDD, Code Review, Agile, Scrum

WORK EXPERIENCE

FieldAssist Dec 2023 – Present · Gurugram, India

SDE-3 (Software Development Engineer III) Promoted from Senior Software Developer, Apr 2026 B2B SaaS Distributor Management System for FMCG brands supporting order management, pricing, invoicing, inven- tory, shipments, and reporting.

• Reduced bulk order processing time from 2 minutes to 30 seconds by redesigning the workflow using NestJS, RabbitMQ, asynchronous processing, and parallel workers.

• Built a proof-of-concept service in Python and FastAPI, working with FastAPI for 6 months alongside the NestJS platform.

• Automated secure report delivery pipelines over SFTP using Node.js and Python for 20+ enterprise clients.

• Improved database and API performance by optimising PostgreSQL queries, fixing N+1 queries, adding indexes, and moving reporting workloads to read replicas.

• Fixed a critical pricing and discount calculation issue that could have resulted in approximately Rs. 10L in incorrect distributor discounts.

• Migrated a 3,000-distributor enterprise client from a shared multi-tenant database to an isolated tenant database for data privacy and scalability, with production validation and rollback support.

• Built a reusable, configuration-driven multi-tenant platform supporting client-specific workflows, feature flags, role- based access, templates, and translations.

• Reduced enterprise client onboarding effort from 2 weeks to 1 week and helped onboard 10+ enterprise clients, including a Coca-Cola bottling partner.

• Fixed concurrent stock allocation issues to prevent multiple requests from claiming the same inventory.

• Led a cross-functional team of 5 engineers across backend, frontend, mobile, integration, and QA; turned require- ments into design docs, mentored junior developers, ran code reviews, and owned the bottling vertical from planning through production launch and support.

Thoughts2Binary May 2021 – Dec 2023 · Gurugram, India Software Engineer Promoted from Associate Software Developer, May 2023 Rollcall — real-time school bus tracking platform serving schools and parents across Australia, New Zealand, and the UK.

• Built and scaled backend services in Python and Node.js for a real-time tracking platform from a 10-school pilot to 200+ schools.

• Reduced database load by approximately 55% and infrastructure costs by 40% by optimising real-time location processing and reducing unnecessary database reads.

• Designed real-time tracking workflows using REST APIs for initial data loading and Socket.IO/WebSockets for live location updates.

• Improved high-volume GPS data processing by moving location history to a datastore better suited for frequent writes and time-series workloads.

• Built secure authentication and authorization with JWT and session management; prevented password-reset abuse of approximately 3,000 requests with CAPTCHA and API rate limiting.

• Built processors and integrations that pull data from external systems, plus reporting APIs, SQL/MongoDB queries and production data repair scripts.

• Led a team of 6 engineers during the absence of the team lead and coordinated development and client support across Australian business hours.

AI PROJECTS

CineMatch AI — Conversational Movie Recommender 2026 Python · FastAPI · React · TypeScript · Supabase · OpenAI API · TMDB API

• Built a full-stack chat app where a FastAPI backend streams LLM tokens to a React + TypeScript UI that renders Markdown as it arrives.

• Stored conversations and messages in Supabase (PostgreSQL) and replayed history to the model for context-aware follow-ups; generated chat titles with a separate LLM call.

• Layered the backend into routes, services and repositories with Pydantic schemas and typed settings; kept it model- agnostic through an OpenAI-compatible endpoint.

• Integrated the TMDB API through tool / function calling and a retrying async HTTP client, so the model fetches live movie data before answering.

RAG Q&A Assistant 2026

Node.js · Gemini Embeddings · Supabase pgvector · LangChain · OpenAI

• Built an end-to-end RAG pipeline: document ingestion, chunking with LangChain’s recursive text splitter, 768- dimension embeddings, and storage in a pgvector vector database.

• Answered questions by embedding the query, retrieving the closest chunks through vector similarity search, and grounding the LLM’s answer in the retrieved context. AI Agent & Multi-Provider LLM Apps 2026

Node.js · OpenAI Responses API · Claude · Gemini · OpenRouter · SQLite

• Built an agentic workflow in which the model calls local tools, gets the results back and decides the next step, with a turn-limit guardrail to stop runaway loops.

• Built an Express + SQLite journaling app with swappable LLM providers (Claude, OpenAI, Gemini, OpenRouter). CERTIFICATIONS

• AI Engineer Path — Scrimba (2026): Agents, RAG, Embeddings & Vector DBs, MCP, Context Engineering, Vercel AI SDK, Open-source Models, Multimodality, Deployment

• Low Level Design (LLD) — Udemy

• Silver Medal in Python & Data Structures — NPTEL EDUCATION

Dr. A.P.J. Abdul Kalam Technical University 2017 – 2021 Bachelor of Technology (B.Tech) in Computer Science 86.4% ACHIEVEMENTS

• Star Performer Award, Thoughts2Binary (2022)

• Pat-on-the-Back Award, FieldAssist (2024)

• Solved 300+ Data Structures and Algorithms problems on LeetCode and GeeksforGeeks

Contact this candidate