MUHAMMAD JAWAD / APPLIED AI DEVELOPER Open to remote opportunities, freelance projects & collaborations

Turning ideas
into intelligent
products.

I build practical AI applications, agentic research systems, Python-powered tools, and automated workflows that turn modern AI capabilities into usable products.

Explore selected work ↘GitHub ↗LinkedIn ↗

Download Resume / CVPDF document

AI APPLICATIONS · AGENTIC SYSTEMS · PYTHON · AUTOMATION

IDEAS → INTELLIGENCEConnected by curiosity.

AI + HUMAN CURIOSITYPYTHON / LLMs / AUTOMATION

BASED IN PAKISTAN · BUILDING FOR EVERYWHERESCROLL TO EXPLORE ↓

01 / SELECTED WORK

Applied intelligence.
Purposeful solutions.

Explore how I bring language models, Python, and thoughtful interfaces together to solve practical problems.

AI VIDEO STUDIOIDEA → VIDEO

01Idea
02Script
03Media
04Voice
05Subtitles
06Render
07Video

SCRIPT · VOICE · SUBTITLESSHORT-FORM VIDEO

FEATURED PROJECT / GENERATIVE AI

AI Video Studio

From a single idea to a short-form video. Automated scripts, voiceovers, stock media, subtitles, and FFmpeg rendering.

PythonFastAPIGroqFFmpegReact

Open app ↗Source code ↗

Inside AI Video Studio

Challenge

Creating a short video requires coordinating scripts, media, narration, captions, and rendering.

Approach

I connected these stages into a single workflow that turns a topic into a short-form video.

Key engineering

FastAPI backend coordinates generation tasks.
Groq generates scripts; Pexels supplies stock media.
Edge TTS provides narration; FFmpeg combines media, audio, and subtitles.

What this demonstrates

AI API integration, media orchestration, and connecting a multi-stage pipeline to a usable interface.

RESEARCHMIND AIAGENTIC RESEARCH

PlanSearchVerifySynthesize

Curiosity,
with direction.

RESEARCH → STRUCTURED REPORTS

02 / AI AGENTS & RAG

ResearchMind AI

A research agent that plans, searches the web, checks information, and writes structured reports with citations.

PythonFastAPIChromaDBTavily

Open app ↗Source code ↗

Hosted on Streamlit. If inactive, the app may need to wake up before loading.

Inside ResearchMind AI

Challenge

Researching a question means gathering information, comparing sources, and organizing findings.

Approach

I built a research workflow that plans tasks, searches the web, processes context, and synthesizes reports with citations.

Key engineering

Research planning and search components coordinate the workflow.
Tavily provides web search; Groq powers language-model calls.
ChromaDB supports document retrieval; fact-checking components review findings.

What this demonstrates

Agent orchestration, retrieval workflows, and structured synthesis. Model-generated findings still require review.

MATHGPTVISION + LANGUAGE

A problem.
A path to understanding.

Image / Text
Understand
Reason
Explain

TEXT OR IMAGE → STEP-BY-STEP EXPLANATIONS

03 / MULTIMODAL AI

MathGPT

A mathematics assistant that reads typed or photographed problems and explains solutions step by step.

PythonGroqVision AIOCR

Open app ↗Source code ↗

Hosted on Streamlit. If inactive, the app may need to wake up before loading.

Inside MathGPT

Challenge

A final answer alone does not explain how to solve a typed or photographed mathematics problem.

Approach

I built a multimodal assistant that accepts text or images and produces step-by-step explanations.

Key engineering

Vision AI processes photographed questions.
Groq provides language-model inference.
OCR components support text extraction from images.

What this demonstrates

Combining image and text inputs with an explanation-focused interface. Generated solutions may need verification.

DOCSHIELDDOCUMENT INTELLIGENCE

File
Identify
Verify
Extract
Inspect

Read deeper.
Verify first.

FILE SIGNATURES · OCR · ANALYSIS

04 / DOCUMENT TOOLS

DocShield

A multi-format document reader with file-signature verification, extension-spoofing detection, OCR, and structured inspection.

PythonStreamlitPyMuPDFOCR

Open app ↗Source code ↗

Hosted on Streamlit. If inactive, the app may need to wake up before loading.

Inside DocShield

Challenge

Documents arrive in different formats, and a filename extension does not reliably identify file content.

Approach

I built a document inspection tool that checks file signatures and extracts content across supported formats.

Key engineering

File-signature checks identify extension mismatches.
PyMuPDF and pdfplumber extract PDF content.
OCR supports scanned text; Streamlit provides the inspection interface.

What this demonstrates

Defensive input handling and multi-format document processing. Signature checks are not malware scanning.

Explore the complete project collection.All repositories ↗

02 / ABOUT & CAPABILITIES

Human curiosity.
Intelligent execution.

I connect ideas, AI models, and practical software to build tools people can use.

THE BUILDER BEHIND THE WORK

Muhammad Jawad

Applied AI Developer · Python & Automation

I focus on applied AI—turning models, APIs, data, and automation into software people can use. My projects explore research agents, document intelligence, multimodal tools, and content workflows.

My work starts with a practical problem: understand the workflow, choose the tools, connect the services, build the application, test the complete flow, and refine the experience.

View résumé

HOW I BUILD

From the problem
to the product.

01
Discover
Understand the problem and desired outcome.
02
Design
Choose the architecture, models, APIs, and workflow.
03
Build
Connect backend logic, AI capabilities, and interface.
04
Test
Validate the user journey and important failure cases.
05
Refine
Improve usability, reliability, and presentation.

Services

FROM YOUR IDEA TO A WORKING APPLICATION

01 / AI APPLICATIONS & AGENTIC WORKFLOWS

Build an AI product
around your workflow.

Research assistants, RAG applications, document intelligence, and agentic workflows that plan, search, process context, and produce structured outputs.

Explore AI projects ↗

02 / WORKFLOW AUTOMATION