Hi, I'm Umar Faaruk — an AI/ML engineer and freelance developer in Hyderabad.
I've shipped 11 client builds: an AI study platform that drew
15,000+ impressions at launch, a live analytics dashboard, a heritage jewellery house,
a Gulf marketplace and more — backed by published deep-learning research.
One developer, end to end — strategy, design, code, deployment and support.
Here's what you can hire me for.
01
Websites, Landing Pages & Rebuilds
Business sites, storefronts and legacy rebuilds that look premium and load fast —
mobile-first, SEO-ready, built to turn visitors into enquiries. I migrated Aerowash off a
legacy WordPress export and shipped Founder & Growth from a blank page.
02
AI & ML Solutions
Custom models, computer vision and LLM features — prototype to production, with the
data pipeline and evaluation to prove they work. Backed by published deep-learning research
(93.5% mAP detection, 99% threat-detection accuracy).
03
Full-Stack Products & Dashboards
Marketplaces, SaaS platforms and live data dashboards with Next.js, TypeScript, FastAPI and
cloud databases — from Edunox and Jobly to the Strategic Arc analytics dashboard that
replaced manual reporting.
04
Deploy, Domains & Support
AWS, Vercel, custom domains and CI/CD via GitHub Actions. Offline-sync for weak networks,
secure API gateways, and fixes after launch — I stick around past the handover.
01 Discover→02 Design→03 Build→04 Launch & Support
// 02. client work
Work That Ships & Sells
11 builds for real clients and products — nine of them live behind the
buttons below, so click through and judge for yourself. The two private builds are an
AI study platform and a client analytics dashboard.
01
AI Product · EdTech
Edunox — AI Study Platform
A self-directed learning platform built and launched solo — React, TypeScript and Vite on a Supabase backend, with Groq-powered AI generating study material on demand. The launch earned 15,000+ LinkedIn impressions and engagement from professionals at Oracle and Morgan Stanley, and ships through a GitHub Actions CI/CD pipeline.
ReactTypeScriptViteSupabaseGroqCI/CD
Independent product — walkthrough available on request.
02
Luxury Retail · Jewellery
Krishna House of Fine Jewellery
Heritage jewellery house with 43+ years of legacy — bespoke 92.5 silver artifacts, bridal collections and rare gems, with WhatsApp ordering built in.
A custom analytics dashboard in Next.js and TypeScript that replaced the firm's manual reporting workflow with a single live, data-driven interface — the numbers now update themselves instead of being assembled by hand.
Next.jsTypeScriptData VizInternal Tool
Client deployment — private to the firm.
05
Luxury Goods · D2C
Raya by the House of Namya
Hand-embroidered 650 GSM cotton towels from Delhi — an editorial storefront built to make a small product feel like a luxury ritual.
A QR table-ordering SaaS built from the ground up — database schema, backend and the customer ordering flow. Guests scan the table code, browse the menu and order from their phone, with no extra hardware for the restaurant.
Doorstep waterless car-wash service, rebuilt from a legacy WordPress export into a fast, modern responsive site — eco positioning, mobile-first layout and a booking flow built for local search.
The engineering depth behind the client work — published deep-learning research,
AI security systems and production LLM builds.
Flagship
Cybersecurity · Deep Learning
Adaptive AI-SIEM for Evolving Cyber Threats
Real-time threat detection across distributed IoT edge nodes. Paired CNN and LSTM models
reached 99% accuracy at a 0.01 false-positive rate on
NSL-KDD and CICIDS2017, with federated learning and blockchain keeping raw data on-device.
Real-time detection pipeline achieving 93.5% mAP at 60 FPS, fit for
production safety monitoring. Built end to end — ingestion, preprocessing,
augmentation, labelling and evaluation — and benchmarked against baseline algorithms
on mAP, F1, precision and recall, with automated Twilio SMS incident alerting.
Accepted for publication in an AI Health Care book (2025, in press).
Conversational AI chatbot built during the Telugu LLM initiative — a Dify.ai
inference API wired into a Streamlit front-end, owning the full data flow from user input
to personalised model response, deployed to a live public environment.
Mobile-first platform deployed on AWS — FastAPI backend, Supabase
PostgreSQL, and a custom offline-sync mechanism with automatic cloud
synchronisation for low-connectivity environments.
AI-powered chess web app — Stockfish engine integrated into a Python
backend for real-time move computation, with an interactive Streamlit UI managing board
state, rule enforcement and outcome detection. Live on Streamlit Community Cloud.
Functional task manager with DOM-driven state management and a mobile-first responsive
layout — one of three JavaScript apps delivered during the Intrainz internship.
I'm Mahmmed Umar Faaruk, an AI/ML engineer and freelance web developer in
Hyderabad, India, and a B.Tech graduate in CSE (AI & ML)
from Sreyas Institute of Engineering and Technology.
I work at both ends of a product. On the front, I ship client builds in
Next.js, React and TypeScript — a live analytics dashboard for
Strategic Arc Consultants, a QR table-ordering SaaS designed from the schema up, and
Edunox, my own AI study platform, whose launch drew
15,000+ LinkedIn impressions and engagement from professionals at Oracle
and Morgan Stanley.
On the back, I build the models. At Viswam AI I contributed to India's first
Telugu LLM initiative alongside Swecha, IIIT Hyderabad and Meta — NLP data pipelines,
prompt engineering and fine-tuning workflows — and architected Craft Connect on
AWS with a secure FastAPI gateway and Supabase PostgreSQL offline-sync.
I'm also a published researcher: my deep-learning work on fire and smoke
detection is accepted for publication in an AI Health Care book (2025, in press), alongside
research in adaptive AI security and real-time computer vision. I like turning complex data
into systems that learn, adapt and ship.
Hyderabad, India
B.Tech CSE-AIML, 2025
Published in AI Health Care (in press)
GitHub
Freelancing — open to projects & roles
Machine Learning
Designing, training and evaluating ML & deep learning models — CNN, LSTM, federated learning.
Computer Vision
Real-time object detection with YOLOv11 & PyTorch — 93.5% mAP at 60 FPS on live video.
GenAI & LLMs
Prompt engineering, LangChain, Groq & Dify.ai — LLM features wired into real products.
Full-Stack Web
Next.js, React & TypeScript front-ends on FastAPI and Supabase back-ends — shipped for paying clients.
AI Security
Adaptive SIEM systems with federated learning & blockchain for evolving cyber threats.
Cloud & Delivery
AWS, Render and Streamlit Cloud with GitHub Actions CI/CD, secure API gateways and offline-sync.
// 05. journey
Experience & Education
Ongoing
Web Developer & AI Engineer · Freelance
Client work end to end — discovery, design, build, deployment and support.
Delivered a custom analytics dashboard for Strategic Arc Consultants
in Next.js and TypeScript, replacing manual reporting with a live, data-driven interface.
Built and launched Edunox, a self-directed AI study platform
(React, TypeScript, Vite, Supabase, Groq) that earned 15,000+ LinkedIn
impressions and engagement from professionals at Oracle and Morgan Stanley,
shipped via GitHub Actions CI/CD.
Designed a QR-code table-ordering SaaS from the ground up —
database schema, backend and customer ordering flow — as a client-ready product.
Rebuilt Aero Wash's site, migrating it off a legacy WordPress export
into a modern, performant responsive build.
Currently building a cinematic, scroll-animated site for a yoga studio client with
custom motion design and mobile-first responsiveness.
Next.jsReactTypeScriptSupabaseGroqCI/CD
May 2025 — Jul 2025
AI Developer Intern · Viswam AI
Regional-language AI, from data pipeline to deployed product.
Contributed to India's first Telugu LLM initiative with Swecha,
IIIT Hyderabad and Meta — NLP data pipelines, prompt engineering and fine-tuning workflows.
Built and deployed the Tune Buddy AI chatbot, integrating a Dify.ai
inference API with a Streamlit front-end across the full request-to-response data flow.
Architected Craft Connect on AWS (ap-south-1) and Render behind a secure
FastAPI + HTTPX gateway, keeping credentials and user data off the client.
Implemented an offline-sync mechanism with automatic synchronisation to
Supabase PostgreSQL, extending access in low-connectivity environments.
FastAPIStreamlitLLMAWSSupabase
Oct 2024 — Dec 2024
Web Development Intern · Intrainz
Foundations in shipping browser applications end to end.
Designed and delivered three JavaScript applications — a calculator,
an e-commerce storefront and a task manager — applying DOM manipulation, state management
and mobile-first responsive design.
Built a personal portfolio site from scratch in semantic HTML, CSS and JavaScript,
with full technical documentation.
HTMLCSSJavaScript
2022 — 2025
B.Tech, CSE (AI & ML) · Sreyas Institute of Engineering &
Technology
Specialised in machine learning, deep learning and artificial intelligence.
Graduated July 2025 — Hyderabad, India.
GitGitHub Actions CI/CDREST API DesignUnit TestingTechnical DocumentationSecure API Gateways
Publication
Employing Deep Learning Paradigms for Fire and Smoke Detection
Accepted for publication — AI Health Care book, 2025 (in press). Also author of
technical articles: NumPy Basics: A Practical Guide with Real-World Case Studies and
Mastering Python Functions (2026).
2025
AWS APAC Solution Architecture
Forage Virtual Experience
2025
Goldman Sachs Software Engineering
Forage Virtual Experience
2025
Tata Data Visualization
Forage Job Simulation
2026
Generative AI Workshop
Growth School
2025
Machine Learning Workshop
Innomatics Research Labs
2023
Machine Learning Workshop
IEEE Sreyas
// 07. playground
Talk to the Terminal
umar@portfolio: ~
// 08. contact
Let's Connect
Have a project in mind?
A website, an AI feature or a full product — let's scope it in one call or one email.