UK

Hi, I'm Umar Khan

Software Development Engineer

Get In Touch View My Work
Umar Khan

About Me

Professional Summary

AI/ML and DevOps engineer with production experience building agentic AI systems, real-time WebRTC platforms, and end-to-end CI/CD pipelines. Proven track record of delivering measurable impact — 92% prediction accuracy, 99% latency reduction, and 60% deployment automation — across internships at NextGen Invent and Cinque Networks.

AWS, MongoDB, and GitHub Copilot certified B.Tech CS senior at VIT (8.53 CGPA) seeking full-time roles in AI engineering, backend development, or cloud infrastructure.

Rath, Uttar Pradesh
VIT Vellore (CGPA: 8.53)
Open to Opportunities

Cloud & DevOps

Expert in AWS infrastructure, CI/CD pipelines, and container orchestration with Kubernetes

AI/ML Engineering

Proficient in building GenAI solutions, LangChain frameworks, and predictive models

Full Stack Development

Skilled in React, Node.js, MongoDB, and building scalable web applications

Security & Performance

Implementing RBAC, vulnerability scanning, and optimizing system performance

Experience

SDE Intern

Jan 2026 – Present

Cinque Networks

Remote

  • Real-Time Communication Architecture: Developed a WebRTC-based communication platform supporting bidirectional video/audio calls, P2P file sharing, and screen sharing using Node.js, WebSocket, and Express.
  • WebRTC Signaling & Orchestration: Architected a custom signaling server with unique callType routing (Video, Audio, Screen, Chat) to prevent unnecessary hardware activation and ensure secure, multi-layered session management.
  • Advanced UI/UX Implementation: Engineered a responsive "WhatsApp-style" Picture-in-Picture (PiP) layout featuring draggable video overlays and viewport-relative (vh) CSS to maintain 16:9 aspect ratios across mobile and desktop devices.
  • System Stability & Debugging: Resolved critical connectivity bottlenecks, including stale DOM references and ICE failure edge cases, achieving 100% session persistence and reliability for subsequent peer connections.
  • Cloud Deployment & DevOps: Managed production deployment on AWS EC2 using PM2 for process persistence and configured self-signed SSL certificates to enable secure getUserMedia browser API access.

Junior AI/ML Intern

May 2025 – July 2025

NextGen Invent

OMR Expressway, Chennai

  • GenAI-Enhanced EDA: Engineered an automated data augmentation pipeline using Google Gemini, expanding training datasets by 40% and improving model robustness for high-risk customer profiling via Exploratory Data Analysis (EDA).
  • Predictive Modeling: Architected multi-model ensembles (Decision Trees, Logistic Regression, Neural Networks), achieving 92% prediction accuracy while implementing ethical guardrails for bias mitigation.
  • Collections Framework: Developed an autonomous agent using LangChain and MCP Server logic to handle collections, reducing delinquency response times by 35% through an automated agentic recommendation loop.

Featured Projects

DevOps: Corporate CI/CD Pipeline Automation

Jenkins Kubernetes Docker SonarQube Prometheus
  • Architected a multi-stage CI/CD pipeline using Jenkins and Kubernetes, reducing manual deployment overhead by 60% and accelerating time-to-market by 100% of critical vulnerabilities identified by Trivy.
  • Configured a secure Kubernetes Control Plane/Worker Node architecture, implementing RBAC and Trivy scanning to mitigate 100% of critical vulnerabilities prior to deployment.
  • Leveraged Prometheus and Grafana to identify a resource bottleneck, leading to a 15% optimization in Kubernetes node resource utilization.

Full Stack AI-Driven Ticket Orchestration

Node.js React MongoDB Gemini AI Inngest
  • Engineered a scalable support system using Inngest and Gemini 1.5 Flash, automating 95% of initial triage tasks and implementing regex-based routing to cut moderator response time in half.
  • Deployed automated skill-based routing logic that extracts required skills from ticket descriptions and automatically assigns tickets to moderators with matching expertise using MongoDB regex queries.
  • Orchestrated an event-driven architecture using Inngest for asynchronous processing of ticket analysis, integrated with JSON Web Tokens (JWT) and Role-Based Access Control (RBAC).

IoT Crop Yield Prediction Architecture

Flask Python ML IoT NoSQL CatBoost
  • Architected an event-driven Flask application to ingest and process sensor streams (temperature, humidity, soil nutrients) for instant yield forecasting.
  • Engineered a custom-trained CatBoost model achieving 98% accuracy using a self-compiled 250,000-row dataset optimized for high-volume IoT ingestion.
  • Optimized NoSQL (MongoDB) API integration to handle high-volume data ingestion; reduced inference latency by 99% (from 15s baseline latency to 150ms optimized CatBoost inference).

Data Warehouse & Analytics Platform

SQL Server ETL Star Schema Medallion Architecture SSMS
  • Architected a medallion-architecture data warehouse (Bronze-Silver-Gold layers) integrating multi-source ERP/CRM data, implementing optimized star schema with dimension/fact tables and comprehensive ETL pipelines for analytical reporting.
  • Established data governance framework including naming conventions, surrogate key management, and technical metadata tracking (data lineage), enabling SQL-based analytics for customer behavior and sales trend insights.

S.A.C.O. — SQL Agentic Co-Pilot Orchestrator

Java AWT Robot JavaFX n8n Gemini AI SSMS MS SQL Server
  • Architected a secure, "always-on-top" AI desktop assistant using JavaFX and AWT Robot to establish a bi-directional context bridge with Microsoft SSMS, enabling automated code extraction and injection without native APIs.
  • Engineered an agentic backend workflow via n8n and Gemini 1.5 Flash, implementing a custom sliding-window memory architecture while preventing API rate-limiting bottlenecks.
  • Developed a human-in-the-loop execution protocol allowing the AI to autonomously draft, review, and optimize SQL queries while strictly safeguarding the database from unauthorized mutations.

Cloud Infrastructure Automation — AWS & Terraform

Terraform AWS IAM S3 VPC EC2 YAML
  • Engineered a scalable IAM user management system using Terraform and YAML-driven configuration, automating provisioning for unlimited users with dynamic policy attachments and reducing manual user creation time by 95%.
  • Architected and deployed an automated S3 static website hosting solution with randomized bucket naming, public access policies, and integrated content delivery, enabling zero-downtime deployments and 100% infrastructure-as-code coverage.
  • Designed a production-ready VPC infrastructure with public & private subnet architecture, automated EC2 web server deployment using user data scripts, and comprehensive security group configurations, reducing infrastructure setup time from hours to 5 minutes.

Practice Projects

Alien Invasion — 2D Arcade Game

Python PyGame
  • Developed a fully functional 2D arcade shooter game in Python where the player controls a spaceship to destroy waves of aliens.
  • Implemented game mechanics including collision detection, scoring, progressive difficulty scaling, and lives system.
  • Built using object-oriented programming principles with modular code architecture across multiple classes.

Expense Tracking System

Python FastAPI Streamlit MySQL Pandas Pydantic Pytest
  • Built a full-stack expense management web application with a FastAPI REST backend and an interactive Streamlit frontend, enabling users to log, update, and analyze personal expenses by date, category, and month.
  • Designed a MySQL database schema for persistent storage and implemented analytics features with Pandas for data aggregation and visualization, along with a pytest-based test suite for backend validation.

Medical Data Extraction System

Python FastAPI OpenCV Tesseract OCR pdf2image Regex
  • Built a REST API backend that accepts scanned PDF documents, applies image preprocessing via OpenCV, and extracts structured fields (patient name, medicines, directions, etc.) using Tesseract OCR and regex parsing.
  • Developed a responsive single-page frontend with drag-and-drop file upload, real-time API health monitoring, and dynamic result rendering for two document types (prescriptions and patient records).

Technical Skills

Languages

Java JavaScript Python YAML SQL Bash

Core Concepts

DSA OOP DBMS Operating Systems Networking

Web & Frameworks

JavaFX FastAPI n8n PyTest Databricks React.js Node.js Express.js TailwindCSS HTML5 CSS3 RESTful APIs

Database

MongoDB MySQL

Generative AI

OpenAI API Hugging Face Ollama & LM Studio LangChain MCP Server

DevOps & Automation

AWS Linux Git GitHub/GitLab Docker Ansible Jenkins Kubernetes Terraform Nginx

Excel Dashboards

Interactive Excel dashboards showcasing data analysis, pivot tables, and visualization skills.

Student Performance Dashboard preview

Student Performance Dashboard

Two-sheet workbook with a raw data layer feeding into a fully interactive dashboard — tracking student scores, attendance, and grade distributions using pivot-driven charts and slicers.

Pivot Tables Dynamic Charts Slicers
Download
Employee Efficiency & Performance Analytics preview

Employee Efficiency & Performance Analytics

Five-sheet workbook analysing workforce productivity — covering break time patterns, efficiency rates, overall performance scoring, and a consolidated KPI summary dashboard.

KPI Tracking Multi-Sheet Analysis Performance Metrics
Download
GETPIVOTDATA Function Mastery preview

GETPIVOTDATA Function Mastery

Advanced practice workbook demonstrating how to use GETPIVOTDATA to pull precise values out of pivot tables for flexible, formula-driven reporting without manual cell references.

GETPIVOTDATA Pivot Tables Formula Reporting
Download
Personal Budget Tracker preview

Personal Budget Tracker

Chapter 2 exercise solution building a monthly personal budget planner with projected vs. actual income and expense categories, variance calculations, and running totals.

Budget Planning Variance Analysis Data Validation
Download

Power BI Dashboards

Interactive Power BI dashboards leveraging DAX, Power Query, and advanced data modeling for actionable insights.

India Election Analysis Dashboard

India Election Analysis (2014–2019)

Comprehensive visual analysis of the Indian General Elections — tracking BJP's growth from 282 to 303 seats, voter turnout surging from 554M to 614M, and deep dives into regional performance, winning margins, and party-wise vote shares.

Power BI Power Query DAX Data Modeling
Download .pbix

n8n Workflows

Automation workflows built with n8n — connecting AI agents, webhooks, and databases for intelligent orchestration.

S.A.C.O. SQL Agentic Co-Pilot Orchestrator

S.A.C.O. — SQL Agentic Co-Pilot Orchestrator

A context-aware desktop assistant bridging AI and SSMS — scans active editor code, sends it to an n8n-hosted Gemini AI agent for debugging/optimization, and injects corrected SQL back into the IDE automatically.

n8n Gemini AI JavaFX Webhooks SQL Server
Download JSON

Education & Certifications

Vellore Institute of Technology (VIT)

B.Tech in Computer Science

2022 – 2026 | Vellore, TN | CGPA: 8.53

Certifications

AWS Certified Cloud Practitioner
MongoDB Database Administrator
Microsoft GitHub Copilot
Associate AI Engineer (Datacamp)
Oracle Cloud Infrastructure 2025 Certified DevOps Professional
GitHub Foundations
Salesforce Certified Agentforce Specialist
Introduction to MongoDB
SQL (Advanced) - HackerRank
Python Data Associate - Datacamp
FastAPI for Machine Learning - CampusX

Get In Touch

Let's Connect

I'm always open to discussing new projects, creative ideas, or opportunities to be part of your visions. Feel free to reach out!

Location

Rath, Uttar Pradesh, India