Javeria Zahid

AI/ML Developer, Data Scientist, and Full-Stack Engineer

Based in: Lahore, PakistanCurrently: AI and ML Trainee, NETSOL Technologies

Computer Science undergraduate with extensive experience building end-to-end AI solutions, including RAG-powered voice assistants and predictive modeling pipelines. Proficient in Python, PyTorch, and the MERN stack, she bridges the gap between machine learning research and scalable full-stack application development.

Experience

AI and ML Trainee

NETSOL TechnologiesinternshipPresent

Summary

Developing Startup Saathi, an AI-powered multi-agent platform for business ideation, compliance guidance, document generation, and strategy support. Working with RAG pipelines, vector search, asynchronous agent orchestration, workflow automation, Deep Learning, Generative AI, Qwen fine-tuning, and LoRA. Applied Python, Pandas, NumPy, Excel, and Power BI to analytics, predictive modeling, and dashboard development.

What I did

Handled multi-language intake and OCR document parsing at scale.

Results

Achieved 98 percent grounding for tax compliance through the RAG pipeline.
Built a house price prediction model using linear regression with custom feature engineering.
Developed a credit score classification system using robust data cleaning and feature selection to achieve high classification accuracy.
Achieved near instantaneous agent response times using Groq and Onnx quantization.

PythonPandasNumPyExcelPower BIRAGVector SearchGenerative AIQwenLoRADeep LearningGroqOnnxOCRscikit-learn

Software Engineer Intern

UDEVSinternship

Built MERN-stack web applications, including e-commerce and marketplace projects, using React, Node.js, Express.js, and PostgreSQL. Developed responsive React interfaces and REST APIs, implemented JWT authentication, managed PostgreSQL data with pgAdmin, deployed projects on Vercel, and maintained code with GitHub.

ReactNode.jsExpress.jsPostgreSQLJWTpgAdminVercelGitHub

Skills

Technical

Python

CSS

HTML

FastAPI

React

JavaScript

NumPy

PyTorch

Scikit-learn

Deep Learning

Generative AI

RAG

Tailwind CSS

Pandas

Excel

Git

React Native

Power BI

MongoDB Atlas

PostgreSQL

LangChain

Computer Vision

XGBoost

LightGBM

Express.js

Node.js

Docker

Groq

AWS EC2

TypeScript

Onnx

OCR

Gemini Live API

WebSockets

Data Augmentation

Label Smoothing

Languages

Urdu

Punjabi

English

Projects

Lahore Air Quality and Weather Analysis

Developeracademic

Data-science project analyzing air quality, weather relationships, and next-day AQI predictions. Analyzed 50K+ AQI rows and 43K+ weather records and built next-24-hour AQI models achieving R² = 0.70.

PythonPandasRandom ForestXGBoost

Brain Tumour Detector

Developeracademic

MRI image classification project for identifying Glioma, Meningioma, Pituitary, and No Tumour classes. Evaluated Random Forest, SVM, and KNN classifiers and achieved 93.4% classification accuracy.

PythonScikit-learnRandom ForestSVMKNN

FarmKonnect: AI Agricultural Marketplace

Developerpersonal

Digital agriculture platform combining marketplace workflows with ML-based crop-price forecasting. Developed a LightGBM forecasting pipeline using web-scraped agricultural market data, feature engineering, and model-drift monitoring. Built FastAPI services and a React Native mobile app using MongoDB Atlas, Docker, and GitHub Actions CI/CD workflows.

LightGBMFastAPIReact NativeMongoDB AtlasDockerGitHub ActionsAWS EC2TypeScriptViteTailwind CSS

Multilingual University Admissions Voice Agent

Developerpersonal

Real-time conversational voice agent that answers university admissions questions through natural speech in English, Urdu, and Punjabi. Used Gemini Live API and FastAPI WebSockets to support bidirectional, low-latency speech-to-speech conversations between the browser client and backend. Built a RAG pipeline using Gemini Embeddings and a custom NumPy cosine-similarity vector store to retrieve accurate and grounded admissions information. Added function calling for dynamic admissions-data retrieval and developed a JavaScript interface using the Web Audio API for microphone capture, audio streaming, and browser playback.

PythonFastAPIGemini Live APIWebSocketsRAGGemini EmbeddingsNumPyJavaScriptWeb Audio API

Breast Cancer Histopathology Classification

Developeracademic

Deep-learning computer-vision project for binary classification of histopathology images. Built a scalable training pipeline using Hugging Face streaming datasets and benchmarked EfficientNet, ResNet-50, and ViT-B/16 architectures. Applied Label Smoothing, AdamW, OneCycleLR, early stopping, and image augmentation to improve model generalization and training stability.

PyTorchHugging FaceAlbumentationsEfficientNetResNet-50ViT-B/16

Startup Saathi

personal

A startup project built using React, Vite, TypeScript, and Tailwind CSS to ensure type safety and scalable component architecture.

TypeScriptReactViteTailwind CSS

Education

Bachelor of Science in Computer Science (BSCS)

University of Management and Technology, LahoreComputer ScienceNov 2022 – Mar 2027

Coursework

Artificial IntelligenceMachine LearningDeep LearningData ScienceData Structures and AlgorithmsDatabasesWeb TechnologiesSoftware EngineeringData Analytics

Hire JaveriaReach out about a role, a contract or a conversation.For recruiters

Javeria's twin is AI, it can make mistakes.

More people on Kanvis

Browse more people Create your own Kanvis