⬡ AVAILABLE FOR SENIOR AI / ENGINEERING ROLES

Satish Kumar

AI Architect & Generative AI Engineer  ·  Fullstack Lead  ·  Noida, India

12+ years designing production-grade Agentic Workflows, RAG Pipelines, and multi-turn LLM systems. Deep hands-on expertise in LangChain, LangGraph, OpenAI, Claude API, Vector Databases — currently driving GenAI adoption for global e-commerce platforms serving Nike across international markets.

12+ Years Exp.
10K+ Daily AI Sessions
35% LLM Cost Reduction
100K+ SKUs in RAG
View Projects Get In Touch
9599745212
satish.gangwar707@gmail.com
https://www.linkedin.com/in/satish-gangwar/
SK
Who I Am

AI Architect Building Intelligent Systems at Scale

I'm a Technical Lead and AI Architect with 12+ years of experience at the intersection of generative AI and fullstack engineering. Currently at Gspann Technologies, I lead the design of production-grade AI systems powering global e-commerce platforms for Nike across multiple international markets.

My core focus is Agentic AI — designing multi-agent orchestration systems, ReAct-based reasoning loops, and RAG pipelines that handle real workloads at scale. I combine deep LLM expertise (LangChain, LangGraph, OpenAI, Claude API) with battle-tested fullstack foundations in React, Next.js, Node.js, and AWS serverless.

Beyond building, I define AI system architecture from greenfield, mentor engineering teams on modern AI/ML patterns, and drive measurable business impact — from 35% LLM cost reduction to 60% latency improvements.

Agentic Systems Design Multi-agent orchestration, ReAct loops, tool-calling, LangGraph workflows
RAG & LLM Engineering Vector DBs, embeddings, retrieval pipelines, LLM guardrails, memory layers
Fullstack Architecture React/Next.js, NestJS, GraphQL, AWS serverless at production scale
Technical Leadership Greenfield architecture, team mentoring, cross-functional delivery
12+
Years of Engineering Experience
5+
Engineers Mentored
~60%
RAG Search Latency Reduction
Sub-2s
AI Agent p95 Latency
10K+
Daily AI Sessions Handled
~40%
Manual Ops Reduction via AI
Technical Expertise

Skills & Technologies

🤖
GenAI & Agentic AI
LangChain LangGraph RAG Pipelines Agentic Workflows Multi-Agent Orchestration ReAct / Plan-Execute OpenAI API Claude API Prompt Engineering Tool / Function Calling Memory Management (ST/LT) LLM Guardrails & Safety
🧠
Vectors & ML
Pinecone (Vector DB) HuggingFace Embeddings OpenAI CLIP Semantic Search Similarity Search Amazon Bedrock Token Budgeting Embedding Optimisation
⚡
Frontend
React.js Next.js TypeScript Redux GraphQL (Apollo) Micro-Frontend Architecture TailwindCSS HTML5 Webpack Jest / RTL
🔧
Backend
Node.js NestJS Express.js Python FastAPI GraphQL (Apollo Server) REST APIs Redis Pub/Sub
☁️
Cloud & DevOps
AWS Lambda API Gateway Amazon Bedrock S3 / SQS / SNS CloudWatch Route 53 Serverless Architecture CI/CD / Jenkins
🗄️
Databases
PostgreSQL Pinecone MongoDB Redis SQL Server Pandas Contentful CMS
Career Timeline

Work Experience

Technical Lead
Gspann Technologies Pvt Ltd
Aug 2021 – Present
React Next.js Node.js NestJS GraphQL AWS LangChain LangGraph GenAI Vector DB RAG PostgreSQL Redis TypeScript
Architected and deployed production-grade Agentic AI systems — LangGraph ReAct workflows, multi-agent orchestration, and LLM guardrails for Nike global e-commerce platforms.
Built a 4-step agentic shopping assistant (intent → retrieval → ranking → generation) handling 10,000+ sessions/day at sub-2s p95 latency.
Engineered RAG pipelines ingesting 100K+ SKUs into Pinecone with HuggingFace embeddings, achieving ~60% reduction in search latency vs keyword-based lookup.
Implemented token-budgeting and response-caching strategies reducing LLM API costs by ~35% while maintaining sub-2s response latency.
Architected headless e-commerce storefront on Next.js with pluggable multi-CMS integration (Salesforce CC, Kibo, Shopify), enabling provider switching with zero frontend disruption.
Led architecture sessions, defined reusable AI component strategy (prompt templates, shared retrieval layer), and mentored 5 engineers.
Sr. Software Engineer
Dentsu World Services
Mar 2021 – Jul 2021
React Node.js GraphQL Okta Redis Pub/Sub AWS S3
Sole architect and lead engineer for Content Symphony — designed entire fullstack architecture from greenfield (React/Redux + Node.js/GraphQL).
Integrated Okta enterprise authentication and built real-time features using Redis Pub/Sub for live notifications between brands and creative agencies.
Associate Consultant
GlobalLogic
Mar 2019 – Mar 2021
React Node.js Micro-Frontend TypeScript Redux
Contributed to Micro-Frontend architecture for SIEM enterprise platform (LogRhythm), enabling independent deployment of feature modules within a unified shell.
Built custom rendering framework to dynamically load and mount Micro-Frontend apps with shared Redux state. Designed metadata-driven UI rendering for Allocate's Health Roster.
Sr. Software Engineer
Acidaes Solutions Pvt Ltd
Jul 2016 – Feb 2019
React Redux REST APIs Node.js Jenkins
Led strategic migration of legacy C# MVC frontend to modern React.js SPA for CRMNext — established new architecture and tooling from scratch.
Configured CI/CD pipelines using Jenkins; mentored developers transitioning from C# to JavaScript/React.
Software Developer
EasyPcCare Pvt Ltd & FinProd Technologies
Jul 2013 – Jun 2016
C# MVC SQL Server ASP.NET jQuery
Developed multi-module university ERP and a BPO-focused CRM system with RDLC performance reports and SSRS executive dashboards.
Project Portfolio

Featured Projects

01 · 2024–PRESENT
Auvira — AI Shopping Platform
Client: GSPANN | Agentic AI · RAG · Multimodal

End-to-end GenAI-powered e-commerce platform featuring a LangGraph-based agentic shopping assistant, multimodal image search, and a headless storefront with pluggable CMS providers.

LangGraph ReAct agentic loop — 4-step workflow (intent → retrieval → ranking → generation) handling 10,000+ sessions/day at sub-2s p95 latency.
RAG pipeline: 100K+ SKUs in Pinecone with HuggingFace embeddings and 512-token chunking — ~60% search latency reduction vs keyword search.
Image-based product search via OpenAI CLIP + Vector DB similarity — ~60% latency improvement over prior text-tag approach.
Token-budgeting and response-caching strategies cut LLM API costs by ~35% while maintaining sub-2s response latency.
Hybrid memory: Redis short-term + PostgreSQL long-term for personalised, GDPR-compliant recommendations.
LangGraph LangChain OpenAI CLIP Pinecone Python · FastAPI Next.js · React PostgreSQL Redis AWS Lambda
02 · 2023–2024
Omni Channel Product Platform
Client: Nike | AI Enrichment · Semantic Search · Serverless

Serverless omni-channel commerce platform for Nike with AI-powered product enrichment, semantic search across 50K+ SKUs, and an agentic validation workflow for merchant feed quality.

AI enrichment pipeline using LangChain + OpenAI API — auto-generated product descriptions, SEO tags, and categories, reducing manual ops effort by ~40%.
Semantic product search using HuggingFace embeddings + Pinecone across 50K+ SKUs, replacing exact keyword matching.
LangGraph agentic validation workflow auto-reviewing merchant feeds for anomalies, duplicate SKUs, and missing attributes with SNS fan-out to human reviewers.
Serverless Lambda + API Gateway architecture processing merchant data at scale with zero server management overhead.
LangChain LangGraph Amazon Bedrock Pinecone NestJS · Node.js AWS Lambda / SQS / SNS React · Next.js PostgreSQL
03 · 2022–2023
C4C Dot Com
Client: Nike | E-Commerce · SSR · Component Library

Nike's consumer-facing e-commerce experience with a reusable UI component library, GraphQL API layer, and SEO-optimised SSR pages.

Designed and published a reusable UI component library reducing duplication and accelerating feature delivery.
Built GraphQL API layer (Node.js + Apollo Server) handling complex data-fetching across multiple backends.
Developed SEO-optimised SSR pages using Next.js, TailwindCSS, TypeScript with full unit test coverage.
React.js · Next.js Node.js · GraphQL TailwindCSS AWS Lambda TypeScript
04 · 2021–2022
GSPANN.COM
Client: GSPANN | CMS-driven · NestJS · Headless

CMS-driven marketing website enabling content team to publish updates via Contentful with zero code deployments, built on a modular NestJS REST API.

Headless CMS integration with Contentful enabling zero-deploy content publishing.
Clean modular NestJS architecture with reusable React component library using Redux and GraphQL integration patterns.
NestJS · Node.js React.js · Next.js Contentful CMS MongoDB TailwindCSS
05 · 2021
Content Symphony
Client: Dentsu | Real-time · Greenfield Architecture

Sole architect and lead for a greenfield fullstack platform connecting brands and creative agencies with real-time collaboration, Okta enterprise auth, and global AWS delivery.

Designed entire fullstack architecture from scratch — React/Redux frontend + Node.js/GraphQL backend.
Real-time notifications between brands and agencies using Redis Pub/Sub.
React.js · Redux Node.js · GraphQL Okta Auth Redis Pub/Sub AWS S3 · Route 53
06 · 2019–2021
SIEM Enterprise Platform & Health Roster
Clients: LogRhythm & Allocate | Micro-Frontend · Metadata-driven UI

Micro-Frontend shell architecture for a cybersecurity SIEM platform and a metadata-driven dynamic UI framework for NHS healthcare scheduling software.

Contributed to Micro-Frontend architecture enabling independent deployment of feature modules within a unified shell with shared Redux state.
Designed metadata-driven UI rendering framework eliminating hard-coded screen layouts for healthcare scheduling.
React.js · TypeScript Redux · Node.js Micro-Frontend Redis · JWT Express.js
Academic Background

Education

2012
B.Tech — Computer Science & Engineering
AIMT, Greater Noida (GBTU)
Score: 63.08%
2007
Senior Secondary (XII)
SRM Inter College, UP Board
Score: 74.00%
2005
Secondary (X)
SVM Inter College, UP Board
Score: 69.83%
Get In Touch

Let's Build Something Intelligent

I'm open to discussing senior AI architect, lead engineer, or GenAI consulting opportunities — especially roles involving Agentic AI systems, LLM-powered platforms, and production-grade RAG architectures.

Phone 9599745212
Email satish.gangwar707@gmail.com
Portfolio https://satish-kumar.pages.dev/
LinkedIn Connect on LinkedIn
Open to Opportunities Based in Noida, India — available for remote, hybrid, or on-site senior AI engineering and architecture roles globally. Experienced delivering for international clients across APAC, EMEA, and North America.
What I'm looking for
Senior AI / Agentic AI Architect roles
GenAI Platform Engineering (LLMs, RAG, Agents)
Technical Lead / Principal Engineer positions
AI-powered fullstack product development