CYBERAKHIL
000%
Home About Skills Projects Journey Contact
NIKHIL
GUMASTA
AI ENGINEER
// CyberAkhil · SATI Vidisha
2nd year B.Tech in AI & Data Science. Building intelligent systems — from LLMs to agentic pipelines & cloud infra.
Scroll
01
About Me

AI Builder.
CyberAkhil.
SATI Vidisha.

Hey — I'm Nikhil Gumasta, known online as CyberAkhil. I'm a 2nd-year B.Tech student in AI & Data Science at SATI Vidisha, and I build real systems — not just coursework.

From self-hosted AI agent infrastructure on Azure VMs to LLM fine-tuning and full-stack applications, I live where theory meets production. My current obsession: agentic AI — autonomous systems that plan, act, and self-correct.

I'm driven by a deep curiosity about intelligence itself — what makes machines understand, and how to build systems that genuinely reason. Beyond code, I explore philosophy, theology, and the meaning of consciousness.

SATI Vidisha Batch 2024–28 AI & DS Azure Cloud LLM Engineer Open to Internships
Nikhil Gumasta
AI & Data Science Engineer · CyberAkhil
// Quick Stats
4+
Projects
125
Samples
6+
Stacks

Skills
& Stack

6 core domains spanning AI engineering, cloud infrastructure, and full-stack development.

01
🧠
AI / Machine Learning
LLM fine-tuning, prompt engineering, RAG pipelines, transformer architectures, and AI APIs at production scale.
LLM APIsFine-tuningRAGChromaDBPrompting
02
⚙️
Agentic Systems
Autonomous multi-agent pipelines — planning, tool use, memory, and task delegation via modern orchestration frameworks.
LangChainLangGraphCrewAIAutoGenOpenClaw
03
🐍
Python & Backend
FastAPI services, data pipelines, JSONL engineering, async patterns, and REST API design from scratch.
PythonFastAPINode.jsREST APIs
04
☁️
Cloud & DevOps
Self-hosted deployments on Azure VMs, Ubuntu server management, process orchestration, and Linux CLI mastery.
AzureUbuntu 24.04Linux CLIGitHub
05
⚛️
Frontend & Full Stack
React interfaces for AI apps, connecting to FastAPI backends with real-time features and modern UX patterns.
ReactJavaScriptHTML/CSSUI/UX
06
☕
Java & DSA
Data structures, algorithms, OOP design patterns, and competitive problem-solving foundations in Java.
JavaDSAOOPAlgorithms

Featured
Projects

04 Works
🤖
// 01 · Featured
OpenClaw AI Agent Platform
  • Full self-hosted AI agent infrastructure on Azure VM.
  • Telegram bot, Gmail via Himalaya, Twilio voice calls.
  • Multi-LLM routing via OpenRouter.
  • Production Linux deployment from scratch.
OpenClawAzure VMNode.jsTelegramTwilioOpenRouter
💬
// 02
Claude-like AI Chatbot
  • Full-stack Claude-inspired assistant.
  • Gemini 1.5 Flash backend, FastAPI, React frontend.
  • RAG via ChromaDB and LangChain.
  • Designed to evolve into a full agentic system.
Gemini 1.5FastAPIReactChromaDB
📊
// 03
LLM Fine-Tuning Dataset
Python script generating 125-sample JSONL training data for Claude-like LLMs. Covers reasoning, Java/DSA coding, structured output & safety — zero validation errors.
PythonJSONLFine-tuning
🔗
// 04 · In Progress
Multi-Agent Agentic System
Evolving the chatbot into a multi-agent autonomous pipeline with LangGraph + CrewAI. Telegram bot as control interface. Self-planning AI with memory & tool-use.
LangGraphCrewAIAutoGen
Drag to explore

My
Journey
So Far.

From enrollment to self-hosted AI infrastructure in under a year — this is the build log.

Let's collaborate →
2024
B.Tech AI & Data Science · SATI Vidisha
Enrolled in AIADS (Batch 2024–2028, #0108AI241045). Building deep ML foundations while simultaneously working on real-world AI projects that go far beyond the curriculum.
Machine LearningAlgorithmsSATI
2025
Self-Hosted AI Infrastructure on Azure
Deployed complete AI agent infrastructure from scratch on Azure for Students VM — nvm, Node.js, OpenClaw v2026.3.23-2, gateway config, port orchestration & multi-service integrations on Ubuntu 24.04.
Azure VMLinuxOpenClaw
2025
LLM Fine-Tuning Dataset Engineering
Designed and generated a 125-sample JSONL fine-tuning dataset covering reasoning, teaching, coding assistance, structured output, and safety — zero validation errors across all records.
JSONLDataset DesignPython
2026
Full-Stack Claude-like AI Chatbot
Architected a production-grade AI assistant: Gemini 1.5 Flash + FastAPI + React + ChromaDB RAG pipeline. Clear roadmap toward full autonomous agentic behavior with LangGraph.
GeminiFastAPIRAG
→ Future
Agentic AI Research & Products
Building toward multi-agent autonomous systems, contributing to AI research, and creating products that genuinely push the frontier of intelligent systems. The mission: build AI that truly understands.
LangGraphCrewAIResearch
Get In Touch

Let's
Build
Something

Open to internships · Research collabs · Interesting conversations about AI

Email GitHub LinkedIn Website