Muhammad Ibrahim

Agentic AI Engineer · Founder & product builder

AI that gets work done.

I build agentic AI workflows that classify, retrieve, decide, and act—from support inbox automation to voice agents. I also ship the APIs, interfaces, and web products that make those systems useful.

Hire IbrahimSee the work
Focus
Agentic AI, retrieval, voice interfaces, and end-to-end automation.
Engineering
Python APIs, full-stack products, reliable data flows, and deployment.
Building
Support inbox agents, Quicklio (35 tools), GradeWise, Kiplio, and The Grid.

Recent work · 2026

AI agents that handle real workflows.

Practical work across support automation, search and retrieval, product engineering, and browser-based tools. Each project includes its actual development status.

Agentic AI / n8n · November 2026Prototype

AI Support Inbox Automation

An end-to-end email support workflow that turns an incoming Gmail message into a classified, tracked, and automatically answered request.

  • A Gemini-powered classifier routes new complaints to a complaint agent and information or complaint-status questions to a separate information agent.
  • The complaint path structures records, assigns a case ID, appends or updates Google Sheets, and drafts a Gmail response.
  • The information path consults a Google Docs knowledge base and case records in Sheets before replying to the sender.
  • Designed to reduce manual inbox triage and repetitive follow-ups. This is a demonstrated workflow, not a claim of production deployment or measured savings.
  • n8n
  • Gemini
  • Gmail
  • Google Sheets
  • Google Docs
  • JavaScript
n8n workflow architecture with Gmail trigger, email classifier, complaint and information agents, Sheets and Docs lookups, JavaScript processing, and automated replies View full size
Architecture diagram redrawn from the actual n8n workflow screenshot (November 2026).
Content automationIn development

Coach Usman Studio

Built for my brother, a gym coach: turn a fitness topic into branded carousel slides and a caption in Roman Urdu or English.

  • n8n content workflow connected to a renderer, with 4/5-slide selection and JPEG/ZIP downloads.
  • Slide-specific word limits, cover variations, and a fixed DM PLAN call to action.
  • Studio backend includes durable jobs and readiness checks for sleeping services; live reliability testing is ongoing.
  • n8n
  • Gemini
  • Node.js
  • PostgreSQL
  • Render
Browser toolsLive

Quicklio

A live library of 35 focused, searchable browser tools for PDFs, images, labels, printing, planning, and everyday file jobs.

  • PDF merge, split, crop, annotation, and image-to-PDF workflows.
  • Image editors, exact-size resizing, print tiling, and practical calculators.
  • Search and category browsing help people discover tools by the job they need done.
  • Supported files are processed locally in the browser; responsive pages and Node.js tests.
  • JavaScript
  • PDF.js
  • pdf-lib
  • Fabric.js
Open QuicklioView code
Illustrated Quicklio tool discovery interface with search and image, PDF, label and planning categories View full size
Illustrative interface graphic; explore the live site for the actual tools.
Retrieval & evaluationInternship project

University FAQ RAG

A university website retrieval pipeline and FastAPI workbench built during my TriNode internship.

  • Crawling, boilerplate removal, content chunking, and cached university indexes.
  • TF-IDF and dense-embedding retrieval with LLM answers and feedback.
  • Evaluated category matching and retrieval ranking on 115 questions.
  • Python
  • FastAPI
  • TF-IDF
  • Embeddings
  • Groq
Conversational agentsPrototype

GeordieDaz

A conversational voice-agent prototype exploring persistent memory and real-time interaction.

  • FastAPI and WebSocket communication with a React interface.
  • Agent orchestration and retrieval of stored facts and conversation history.
  • PostgreSQL/pgvector for memory, with Redis for working context.
  • FastAPI
  • LangGraph
  • WebSockets
  • pgvector
  • React
View code
Education automationPrototype

MCQ Generation & Grading

An internship prototype connecting document-to-MCQ generation with printable answer sheets and grading.

  • Generate questions from source documents and prepare printable sheets.
  • Explore OpenCV/ArUco-based answer-sheet grading and Excel exports.
  • Python
  • LLM APIs
  • OpenCV
  • ArUco
  • Excel
Structured extractionPrototype

Receipt Expense Workflow

An n8n experiment for turning receipt images into structured expense records.

  • Extract brand, line items, prices, total, and date/time into structured output.
  • Prototype scope; extraction accuracy and downstream integrations remain under development.
  • n8n
  • Document extraction
  • Structured output

Experience

AI work, with engineering foundations.

Founder-led product work alongside internships in agentic AI, security operations, and data science.

Independent ventures · Ongoing

Leading products, not just writing code.

CEO / founder of Kiplio and The Grid, and the developer behind Quicklio. My current work connects AI workflows with usable web products—from knowledge-backed agents to publishing and practical browser tools. These are independently built initiatives; I do not present prototype work as enterprise-scale deployments.

  • Kiplio · Founder / CEO
  • The Grid · Founder / CEO
  • Quicklio · Product development

  1. 16 July – 9 September 2026

    TriNode

    Agentic AI Engineer Intern

    • Built a university FAQ RAG pipeline with crawling, text cleaning, TF-IDF and dense retrieval; evaluated matching and ranking on 115 questions.
    • Developed a FastAPI workbench with cached indexes, LLM answers, and feedback.
    • Prototyped document-to-MCQ generation, printable sheets, OpenCV/ArUco grading, and Excel exports.

  2. March – June 2026

    SecureDevLabs

    Cyber SOC Intern

    • Investigated simulated incidents using Wireshark, Suricata, Zeek, and QRadar.
    • Mapped activity to MITRE ATT&CK and documented malware indicators, detection rules, and a NIST-aligned response plan.

  3. May – July 2023

    CodeAlpha

    Data Science Intern · Remote

    • Built a Power BI Titanic dashboard and Python stock-prediction notebooks.
    • Ran simulated A/B tests with randomized groups and hypothesis testing.

Selected product evidence

Selected builds, end to end.

AI workflows, backend systems, and the interfaces people use to work with them.

GradeWiseWorkbench overview
Real GradeWise product captures from the grading workspace.

01

GradeWise

AI-assisted evaluation that gives structured feedback at scale.

My final-year project: document intake, OCR, course-material retrieval, rubric-based grading, similarity review, and downloadable reports with instructor approval.

Role
Full-Stack / AI Engineering
Stack
Next.js / FastAPI / PostgreSQL / Docker / RAG / Async Workers

What this proves

  • Sole developer across the frontend, backend, grading workflow, and deployment.
  • PDF, DOCX, image, and code submission workflows with PDF/Excel reporting.
  • Deployed demo on Vercel and Render; backend currently paused.
Open Repo
BulaoVoice booking flow
Voice-booking product concept imagery supplied for the Bulao workflow.

02

Bulao

Voice-first booking built for real conversations in Urdu and English.

A six-stage Gemini backend for Urdu and English service booking: intent detection, provider discovery and ranking, PKR estimates, booking, and follow-up.

Role
Backend / AI Workflow Engineering
Stack
Python / FastAPI / Gemini / Agentic AI / Docker / Cloud Run

What this proves

  • Natural-language requests mapped into structured service-booking actions.
  • FastAPI backend with Gemini integration, pytest, and Docker.
  • Urdu and English conversation flows with explicit stages and API boundaries.
Repo
The GridNode homepage
Real local capture from the preserved /node route.

03

The Grid

A publishing platform for creating, curating, and connecting around ideas.

A full-stack content platform with rich-text authoring, authentication, email flows, and a deliberate community experience.

Role
Full-Stack Engineering
Stack
Next.js / Supabase / PostgreSQL / Authentication / Email

What this proves

  • Preserved local route with publishing, identity, and content workflows.
  • Shows product thinking beyond portfolio pages: platform, posts, and community.
  • Full-stack surface connected to database, auth, and email flows.
Open

Engineering depth

Under the product layer: systems, algorithms, and low-level work.

The portfolio leads with shipped interfaces, then backs them with the foundations teams expect from a serious full-stack engineer.

Red-Black Tree

Templated balancing, insertion, deletion, traversal, and tree rotation implemented from first principles.

insert()
rotateLeft()
fixViolation()
levelOrder()
Inspect repo

CPU Scheduling

FCFS, SJF, and Round Robin process simulation with turnaround and waiting-time output.

P1 | P2 | P3
Gantt chart
turnaround
waiting time
Inspect repo

x86 Assembly

Registers, memory, stack, procedures, interrupts, loops, arithmetic, and low-level logic.

MOV AX
INT 21h
stack
procedure
Inspect repo

How I fit into a product team

Three operating modes, one engineering profile.

Use me where product surface, AI workflow, and engineering foundation meet.

01

Product Systems

Interfaces, dashboards, content platforms, ecommerce, authentication, and user workflows.

Next.js / React / TypeScript / Tailwind / Shopify / Supabase

02

Agentic AI & Automation

Specialized agents, classified inboxes, case tracking, LLM integrations, knowledge retrieval, and practical automation.

n8n / Gemini / Gmail / Google Sheets / FastAPI / RAG / pgvector / LangGraph

03

Engineering Foundations

Backend architecture, APIs, databases, deployment, algorithms, and systems thinking.

Python / C++ / PostgreSQL / Docker / Linux / x86 Assembly

Verification

Credentials, education, and contactable proof.

Structured signals for teams that need to validate fit before a call.

Credentials
CredentialIssuerYear
Agentic AIDeepLearning.AI2025
Certified in CybersecurityISC22026
API Fundamentals Student ExpertPostman2024
Object-Oriented Programming in C++Coursera-
Programming for EverybodyCoursera-
Introduction to SQLDataCamp2024

Available for AI engineering and automation work

Have a workflow that could use an AI agent or better automation?

Send the brief
EmailDirect briefLinkedInProfessional profileGitHubCode proofThe Grid /nodeLive platform