01
Machine Learning
Designing supervised pipelines end-to-end — feature engineering, class balancing (SMOTE), PCA, model selection between RF / XGBoost, and rigorous evaluation (ROC-AUC, F1, latency).
scikit-learn · XGBoost · Pandas
Open to AI / ML, Full Stack & MERN rolesEngineering intelligent systems
Engineering intelligent systems
from data → decisions.
I'm Numair Fahad — a Computer Science student training as an AI / ML Engineer, with strong experience as a Full Stack & MERN Developer. I build production ML pipelines, ship full-stack products powered by LLMs, and engineer secure systems end-to-end.
$ currently learning
Learning Python programming in advance and depth — advanced OOP, decorators, generators, asyncio concurrency, typing with Pydantic, profiling, and pytest — so my ML code ships production-grade.
more on my About page →AI / ML skills at a glance
Pythonscikit-learnXGBoostPyTorch-readyNode.jsReactDocker
98.63%
Test accuracy — VPN traffic model
6+
Shipped projects across AI, full-stack, security
20+
Tools & frameworks in active use
2
Internships — engineering & data
Three lanes I'm pushing across AI & full-stack engineering.
02
Applied AI Products
Wiring LLMs into real apps — Gemini and OpenAI APIs for resume review, JD generation, and intelligent workflows behind authenticated MERN dashboards.
Gemini · OpenAI · Node · React
03
Secure Systems
Background in cybersecurity & networking — Wireshark, Nmap, OWASP ZAP, Linux. Useful for building AI systems that handle data and traffic responsibly.
Linux · Wireshark · Cisco
AI / ML highlights
All projects →Featured AI & data projects
A mix of production ML pipelines, applied AI products, and data foundations — each with clear impact, stack, and next steps.
Classical ML · Cybersecurity/vpn-ml
AI-Enhanced VPN Traffic Classifier
Encrypted-traffic ML pipeline that classifies VPN vs non-VPN flows using only metadata — no payload inspection.
98.63%
Test accuracy on the ISCX VPN-nonVPN dataset
Pythonscikit-learnXGBoostSMOTEPCAPandasNumPyMatplotlib
Generative AI · RAG/learntube
LearnTube — AI YouTube Learning Assistant
RAG application that converts any YouTube video into a grounded chat, summary, quiz, and flashcards — on web and native mobile.
~85%
Query intent-classification accuracy routing RAG skills
PythonFastAPIGoogle GeminiChromaDBsentence-transformersscikit-learnReactExpo
Applied AI · Full-Stack/smarthire
SmartHire AI — Job Portal
MERN job portal with AI resume review and AI-generated job descriptions, built for students, recruiters, and admins.
3 roles
End-to-end dashboards with AI-powered workflows
ReactRedux ToolkitNode.jsExpressMongoDBSocket.IOGemini APIOpenAI API
Data Engineering · ML Prep/clinical-data-pipeline
Clinical Data QA & Reporting Pipeline
Reproducible data-processing workflow for a clinical services team — cleaning, validation, and reporting at scale.
↑ accuracy
Improved reporting accuracy and turnaround time
PythonSQLExcelData QAReportingETL
Featured work
All projects →Selected projects
Filter by discipline or tech to find what's most relevant to you.
Category
Technology
Work context
7 of 7 project(s)
AI / ML/vpn-ml
onsite
AI-Enhanced VPN Traffic Classifier
National University of Technology (NUTECH)
Encrypted-traffic ML pipeline on the ISCX VPN-nonVPN dataset — 98.63% test accuracy.
- ▸End-to-end ML workflow: preprocessing, scaling, SMOTE/PCA, train/val/test splits.
- ▸Random Forest baseline + XGBoost comparison for real-time VPN gateway inference.
- ▸Evaluated with accuracy, precision/recall, F1, ROC-AUC, feature importance, latency.
Full-Stack/smarthire
remote
SmartHire AI — Job Portal
DawoodTech NextGen
Full-stack MERN platform with AI resume review and AI-generated job descriptions.
- ▸Role-based dashboards (students / recruiters / admins) with JWT auth.
- ▸AI resume review + job description generation via Gemini / OpenAI APIs.
- ▸Realtime notifications with Socket.IO, file uploads via Multer + Cloudinary.
AI / ML/learntube
remote
LearnTube — AI YouTube Learning Assistant
Zynvex Solutions
Gemini + RAG app that turns any YouTube video into a summary, quiz, flashcards, and a transcript-grounded chat.
- ▸Capstone of the 6-week remote AI/ML internship at Zynvex Solutions.
- ▸RAG pipeline: transcript ingestion → semantic chunking with timestamps → sentence-transformers embeddings in ChromaDB.
- ▸TF-IDF + Logistic Regression intent classifier (~85% accuracy) routes each query to the right skill.
Cybersecurity/airport-net
onsite
Smart Airport Network
National University of Technology (NUTECH)
Hybrid wired/wireless campus network simulated in Cisco Packet Tracer.
- ▸Cisco 2960 switches + 2811 routers, WRT300N wireless, WPA2-Personal.
- ▸DHCP, static addressing, NAT/gateway routing across segmented subnets.
- ▸L2 switching + L3 routing across wired, wireless, and router-link networks.
Systems/event-ticketing
onsite
Campus Event Ticketing System
National University of Technology (NUTECH)
Hybrid C++ / Python desktop app — DSA-driven backend with a Tk dashboard.
- ▸Singly linked list events, BST lookup, FIFO walk-in queue, selection sort by date.
- ▸SQLite persistence for events + participants across sessions.
- ▸CustomTkinter dashboard talks to C++ engine via subprocess commands.
Systems/typing-tutor
onsite
Typing Tutor for Novices
Bahria University
Falling-alphabet typing game built in C++ with SFML.
- ▸OOP structure: MainMenu, LoginPage, GameLevel, CreditPage with UML diagram.
- ▸File-based auth, real-time input, random alphabet generation, scoring + penalties.
- ▸Audio, background art, and polished menu/credits flow.
Embedded/iot-car
onsite
IoT Obstacle-Avoiding Robotic Car
Bahria University
ESP8266 NodeMCU prototype with a browser-based control surface.
- ▸Ultrasonic sensing + microcontroller decision logic for obstacle avoidance.
- ▸Local web-server interface to drive GPIO outputs from a phone browser.
- ▸Embedded circuit work on breadboard with LEDs, resistors, ESP8266.
Experience
Full timeline →Where I've been building
Jul 2026 – Aug 2026 · 6 wks
AI / ML Intern
Zynvex Solutions · Remote
scikit-learn · TensorFlow · Gemini · RAG · FastAPI
Apr 2026 – Jun 2026
Full Stack Engineer (Intern)
DawoodTech NextGen · Rawalpindi, PK
MERN · JWT · Socket.IO · Cloudinary
Aug 2024 – Jul 2025
Data Processor & Programmer (Intern)
Health and Psychiatry Services · Tampa, FL
Python · SQL · Data QA · Reporting
The stack I build with
AI / ML & Data
scikit-learnXGBoostTensorFlowKerasPyTorchDeep LearningCNNsRNN / LSTM / GRUPandasNumPyMatplotlibSeabornOpenCVEDAFeature EngineeringTF-IDFLogistic RegressionSMOTEPCACross-ValidationHyperparameter TuningJupyterjoblib
Generative AI
LLMsPrompt EngineeringRAG PipelinesVector SearchEmbeddingssentence-transformersChromaDBHuggingFaceGemini APIGemini Flashgoogle-genai SDKJSON-mode structured outputOpenAI APIONNX RuntimeModel Evaluationn8n AI Agents
Languages
PythonTypeScriptJavaScriptJavaC / C++SQLRISC-V / x86 Assembly
Backend
FastAPIUvicornPydanticNode.jsExpressREST APIsSocket.IOJWT / PyJWTBcryptMulterNodemailerAxiosSQLAlchemySQLiteslowapi (rate limiting)
Frontend
ReactRedux ToolkitTanStack QueryViteTypeScriptTailwind CSSshadcn/uiRadix UIFramer MotionHTML5CSS3
Mobile
ExpoReact NativeExpo RouterClerk AuthiOS / Android buildsMobile UX
Data & Infra
MongoDB AtlasPostgreSQL / SupabaseChromaDBSQLiteSQL ServerDockerKubernetesCloudinaryAzure App ServiceVercelReplit deploypnpm monorepoyt-dlp / youtube-transcript-apiGit & GitHubCI/CD
Security & Systems
Linux / KaliWiresharkNmaptcpdumpOWASP ZAPAutopsyCisco Packet Tracer
Let's build
Looking for an AI / ML or Full-Stack engineer who ships?
I'm open to AI Engineer, ML Engineer, ML Specialist, AI Researcher, Full Stack Developer, and MERN Stack Developer internships and junior roles — remote, hybrid, and onsite-friendly. Let's talk about how I can help your team move from prototype to production.