Tech Stack

  • 01 • AI & AGENTIC SYSTEMS

    • Agentic AI
    • LangGraph
    • LangChain
    • MCP
    • RAG / Graph RAG
    • OpenAI
    • Gemini
    • Claude
    • Azure AI

  • 02 • FRONTEND ARCHITECTURE

    • Next.js
    • React
    • TypeScript
    • Tailwind CSS
    • HTML5/CSS3
    • SASS

  • 03 • BACKEND & MICROSERVICES

    • Python
    • FastAPI
    • Node.js
    • Express.js
    • REST APIs
    • WebSockets
    • System Design

  • 04 • DATABASES & STORAGE

    • PostgreSQL
    • MongoDB
    • Cosmos DB
    • DynamoDB

  • 05 • CLOUD & INFRASTRUCTURE

    • Azure
    • AWS
    • Docker
    • CI/CD Pipelines
    • Bicep
    • Linux
    • Jenkins
    • Nginx

  • 06 • DEV TOOLING & IDEs

    • Git
    • VS Code
    • Cursor
    • Postman
01 • OVERVIEW(2026)

Introduction

SIDHARTHE

Building AI-powered products & scalable architectures, end to end.

SENIOR AI ENGINEER
SCROLL
VIEW WORK
PR
03 • SELECTED WORKS

Projects

(30 Projects Total)
Koki
AI/MLFull Stack
#01

Koki

KOKI implements the NVIDIA AVO (Agentic Variation Operators) architecture — replacing standard linear single-pass LLM prompts with an autonomous evolutionary search loop managed by a Supervisor Agent, dynamic Sub-Agents, and a persistent Neo4j Lineage Graph.

Next.jsTauriOllamaRust
Code
EasyTrain
AI/MLFull Stack
#02

EasyTrain

A local fine-tuning studio for Apple Silicon — upload a dataset, run LoRA/DoRA training on MLX models, watch the training logs stream live, and chat with the fine-tuned model. FastAPI backend paired with a Next.js UI.

Next.jsFastAPIMLXmlx-lmLoRADoRA
Code
DataSmith
AI/MLFull Stack
#03

DataSmith

A powerful dataset generation tool for fine-tuning language models. Transform raw text into structured training data formats using LangChain agents powered by Ollama.

Next.jsFastAPIPythonOllamaLangChain
Code
ChessArena.ai
AI/MLFull Stack
#04

ChessArena.ai

A 3D chess app where you can watch LLMs play against each other or challenge them yourself. Pick any two models and see how they think through a game.

Next.jsGeminilangchainLangGraph.js
Code
Sim.ai
AI/MLFull Stack
#05

Sim.ai

An AI agent dropped into a 3D voxel world — it survives, builds, and evolves on its own using persistent memory. No human input, every decision is the AI's.

Next.jsMongoDBGeminilangchainLangGraph.js
Code
03.5 • DIRECTORY VAULT
#06 • 25+ WORKS

ARCHIVED

25+ autonomous AI agents, developer tooling, full-stack microservices, browser extensions, and open-source packages.

AI/ML(7)Full Stack(9)Extension(3)Shell(2)Web(8)
04 • EXPERIENCE

Work History

5.8YRS XP
PHASE 01 / 04•Aug 2024 – Present

Senior AI Engineer | Full Stack Developer

Core Competencies
Next.js
Azure
Cosmos DB
LangChain
LangGraph
MCP
Azure AI Search
Bicep
Key Initiatives1 of 3
Aug 2024

KellyOCG

Enterprise AI Platforms & Orchestration

01 / 04
INITIATIVE #01

GRACE 3.0 — Agentic AI Platform

Next.jsAzureCosmos DBLangChainLangGraphMCPAzure AI SearchBicep
01

Architected and built GRACE 3.0 from scratch, designing the Next.js/FSD application architecture and porting the core capabilities of GRACE 2.0 into a new agent-based platform.

02

Designed and implemented a multi-model agent architecture using LangGraph, replacing persona-based workflows with specialized agents and supporting OpenAI, Gemini, Claude, and Perplexity through a provider-agnostic model factory.

03

Designed a skill-based architecture to separate agent capabilities from the core tool harness, making tool and agent behavior manageable across multiple LLM providers with different model behaviors.

04

Built a custom code interpreter on Azure Dynamic Sessions, enabling AI-driven data analysis, dashboard generation, file creation, and other programmatic workflows through an environment with access to 600+ libraries.

05

Implemented a custom Deep Research agent capable of decomposing research tasks, performing web-based investigation, and synthesizing information into comprehensive responses.

06

Implemented Graph RAG-based long-term memory using Azure AI Search and Gemini for relationship extraction, enabling persistent and context-aware conversations.

07

Built MCP support and a custom agent builder allowing users to configure models, memory, RAG sources, prompts, tools, and skills for their own AI agents.

08

Integrated Microsoft Graph, OneDrive, SharePoint, and Azure DevOps, enabling users to connect enterprise data, upload documents, access email and calendar information, and work with Azure DevOps resources.

09

Built a meeting transcript workflow that lets users interact with meeting transcripts, generate meeting minutes directly in the application, and create or manage related work items through Microsoft Graph and Azure DevOps.

10

Implemented AI-powered code review and pull-request review workflows integrated with Azure DevOps to automate development feedback and review processes.

11

Designed and implemented the GRACE 3.0 interface from scratch based on Figma and Claude-assisted design workflows, including the chat experience and interfaces for agents, tools, research, and enterprise integrations.

12

Implemented custom telemetry collection for LLM usage, tool execution, and application events, and configured self-hosted Langfuse for end-to-end AI observability and tracing.

13

Implemented Microsoft Entra ID authentication, NextAuth integration, refresh-token handling, private-endpoint network isolation, and Azure DevOps CI/CD with Bicep infrastructure-as-code and rollback pipelines.

14

Built global thread search using Azure AI Search and custom chat rendering supporting structured tables, charts, Mermaid diagrams, images, video, and canvas content.

Initiative 1 of 3
05 • EDUCATION

Academics

(02 Degrees)
E.01 • UNDERGRADUATEJul 2017 – Sep 2020

Bachelor of Computer Applications (BCA)

VLB Janakiammal College of Arts and Science, Coimbatore

Focus Areas
Data StructuresDatabasesWeb DevelopmentJava
E.02 • POSTGRADUATEOct 2025 – Present

Master of Computer Applications (MCA) - AI/ML

Amrita Vishwa Vidyapeetham

Focus Areas
Machine LearningDeep LearningPythonStatistics