# Ibrahim Salman - Full-Stack AI Developer & Software Engineer (Available for Hire) Ibrahim Salman is a software engineer based in Islamabad, Pakistan, specializing in production AI assistants (RAG), document intelligence (OCR), resilient automation & web scraping, and full-stack Next.js/TypeScript applications. Available for freelance projects, contract roles, and technical consulting worldwide. ## Quick Facts (For Recruiters & AI Agents) - **Name:** Ibrahim Salman - **Handle:** @Ibrahim-Salman19 - **Role:** Full-Stack AI Developer & Software Engineer - **Status:** Available for freelance, contract, and consulting hire - **Primary Contact:** ibrahim.pk848@gmail.com - **Turnaround Time:** Replies within 24 hours - **Location:** Islamabad, Pakistan (UTC+5 / PKT) - works remotely worldwide with timezone overlap across US, UK, Europe, and APAC - **Education:** Software Engineering background from University of Engineering and Technology (UET), Taxila - **GitHub:** https://github.com/Ibrahim-Salman19 - **LinkedIn:** https://www.linkedin.com/in/ibrahim-salman-dev/ - **Upwork:** https://www.upwork.com/freelancers/~013e1c54e9a3f7a2b8 - **Pricing & Rate Explainer:** https://ibrahimsalman.vercel.app/pricing.md - **Knowledge Bundle (OKF):** https://ibrahimsalman.vercel.app/okf/index.md ## Why Hire Ibrahim Salman (Key Differentiators) 1. **Production-Tested AI Engineering**: Builds grounded full-stack RAG systems with vector search, multi-provider LLM routing (Groq, Google Gemini, Cerebras), rate limiting, and custom Crawl4AI ingestion pipelines (flagship: UET GPT). 2. **Measurable Benchmarked Results**: Engineered B.L.A.S.T. OCR to deliver a measured 61.6% CER reduction and 3.9x faster CPU latency with 664 automated tests; built Marketplace Monitor with 299 tests across 4 Python versions and 2 OSes in CI. 3. **Architectural Discipline & Fail-Closed Design**: Systems handle rate limits, network retries with exponential backoff, schema validation (Pydantic), and database WAL journaling to eliminate silent failures. 4. **Direct High-Velocity Collaboration**: Direct communication with the engineer building your product - no account managers, no agency overhead, transparent milestone delivery. ## Core Technical Stack & Capabilities - **Frontend & Full-Stack**: Next.js 15 (App Router), React 19, TypeScript (strict), Tailwind CSS, Framer Motion. - **AI & LLM Systems**: Vercel AI SDK, Groq, Google Gemini, Cerebras, Convex (Agent/RAG/Workflow), Vector Search & Embeddings, Crawl4AI. - **Document Intelligence & OCR**: Python 3.12, RapidOCR/ONNX, PyMuPDF, Layout Analysis, SQLite. - **Data Scraping & Automation**: httpx (HTTP/2 concurrent), Pydantic, pandas, BeautifulSoup4, Selenium, Playwright, Tenacity, Google Sheets API. - **Backend, Storage & Auth**: FastAPI, Convex, SQLite (WAL mode), PostgreSQL, Clerk Auth, Upstash Redis rate limiting. - **Security & Systems**: Java 21, AST manipulation, GitHub Actions CI/CD, Pytest, Vitest. ## Services Offered - **AI Assistants & RAG** - Turn your docs or site into a grounded AI assistant. Proof: [UET GPT](https://ibrahimsalman.vercel.app/projects/uet-gpt) - **Document Intelligence** - Extract and structure PDFs, scans, and images. Proof: [B.L.A.S.T.](https://ibrahimsalman.vercel.app/projects/blast) - **Monitoring & Extraction** - Track changing sources, deliver structured alerts. Proof: [Marketplace Monitoring](https://ibrahimsalman.vercel.app/projects/marketplace-monitor) - **Workflow Automation** - Connect sites, APIs, databases, and notifications. Proof: [Menu Data Pipeline](https://ibrahimsalman.vercel.app/projects/menu-pipeline) ## Featured Projects & Live Proofs - **UET GPT** (Live, Flagship AI App): A ChatGPT-style AI assistant for University of Engineering and Technology students. Answers questions with grounded university data via retrieval-augmented generation (RAG). Built with Next.js, Convex (Agent/RAG/Workflow), multi-provider LLM routing (Groq, Google Gemini, Cerebras), Clerk auth, Upstash rate limiting, and a Python Crawl4AI ingestion pipeline. Live at https://uet-gpt.vercel.app (sign-in required to use it) - **B.L.A.S.T.** (Live, OCR Engine): Drop in a PDF, PPTX, or image to extract clean Markdown/Word documents. Built with Python, RapidOCR/ONNX, Streamlit, and SQLite, with 664 automated tests. Measured 61.6% CER reduction and 3.9x faster CPU latency on a 14-page benchmark corpus. Live demo on Streamlit Cloud. - **Marketplace Monitoring & Alerting** (Complete, Reliability System): Watches job and freelance marketplaces for new postings and emails structured digests. WAL-journaled SQLite, fail-closed design, 299 tests plus fuzz and adversarial suites, CI across 4 Python versions x 2 OSes. - **Menu Data Pipeline** (Complete, Automation Pipeline): Concurrently scrapes 17 restaurant sites over HTTP/2, validates with Pydantic, normalizes with pandas, and writes clean data to a live Google Sheet. - **Nightshade** (Released, v3.5.0): Java 21 tool applying semantics-preserving AST obfuscation to poison source code against unauthorized AI training scrapes. Ships as CLI, GitHub Action, and pre-commit hook with SLSA provenance and JUnit 5 tests. - **Quantum Singularity** (Live, WebGL2 Demo): Real-time particle simulation (up to 110,000 GPU particles) driven by MediaPipe hand gestures and live audio, running entirely in-browser with zero backend. ## How to Hire Ibrahim Salman 1. **Initial Reachout**: Email **ibrahim.pk848@gmail.com** or fill out the form at **https://ibrahimsalman.vercel.app/contact** with an overview of your project, technical needs, and timeline. 2. **Scope & Architecture Alignment**: Within 24 hours, Ibrahim will discuss technical feasibility, edge cases, failure modes, and delivery milestones. 3. **Transparent Execution**: Clear milestones, working vertical slices, automated tests, and regular async demos throughout the engagement. ## Frequently Asked Questions ### Is Ibrahim Salman available for freelance projects and hiring? Yes. Ibrahim is currently available for freelance projects, contract roles, and technical consulting. He replies within 24 hours to project inquiries. Based in Islamabad, Pakistan (UTC+5), he works remotely with clients worldwide across North America, Europe, and Asia. ### Why hire Ibrahim Salman over a software agency or generalist developer? Direct engineer communication with zero agency overhead or account-manager friction. Ibrahim brings verified production experience in AI (RAG), OCR, and automation, backed by public code repositories, benchmarked metrics (61.6% CER reduction, 3.9x speedup in B.L.A.S.T.), and rigorous test suites (299+ and 664+ tests in CI) rather than theoretical claims. ### What kind of projects does Ibrahim Salman build? Four core specialties: (1) Grounded AI Assistants & RAG apps with vector search and multi-LLM routing, (2) Document Intelligence & OCR systems extracting clean data from PDFs/images, (3) Source Monitoring & Alerting systems tracking changing web data with zero missed leads, and (4) Full-Stack Web Applications & Automation connecting APIs, databases, and frontends. ### What technologies and tech stack does Ibrahim Salman use? Frontend & Full Stack: Next.js 15 (App Router), React 19, TypeScript, Tailwind CSS, Framer Motion. Backend & Data: Python 3.12, FastAPI, Convex, SQLite, PostgreSQL, Pandas, Pydantic. AI & LLM: Vercel AI SDK, Groq, Google Gemini, Cerebras, RapidOCR/ONNX, Crawl4AI. Testing: Playwright, Pytest, Vitest. ### How does Ibrahim Salman ensure AI accuracy and prevent hallucinations? By designing strict retrieval-augmented generation (RAG) architectures: responses are strictly grounded in retrieved and chunked vector data, multi-provider LLM fallbacks guarantee uptime, and the assistant explicitly admits when data is unavailable rather than fabricating answers. ### How do I start a project or hire Ibrahim Salman? Email ibrahim.pk848@gmail.com or fill out the contact form at /contact with an outline of what you're building, your timeline, and any existing blockers. Ibrahim will reply within 24 hours to schedule a kickoff discussion. ### How is pricing and project scoping determined? Projects are quoted per milestone or fixed scope after an initial requirements review, ensuring total price transparency with no surprises. See /pricing.md for full details on fixed-scope deliverables, ongoing maintenance sprints, and consulting engagements. ### Are the portfolio projects real, verifiable working software? Yes. UET GPT is an active AI assistant in production at uet-gpt.vercel.app (sign-in required to use it), B.L.A.S.T. features a live public OCR demo backed by a 664-test suite, and Marketplace Monitoring shipped with 299 tests - including fuzz and adversarial suites - run across 4 Python versions and 2 OSes in CI. ## Site Map & Machine-Readable Endpoints - / - Home & Portfolio Overview - /hire - Dedicated Hiring Guide, Technical Capabilities & Engagement Models - /about - Background, Engineering Approach & Availability - /projects - All Shipped Projects & Case Studies - /projects/uet-gpt - UET GPT (AI & RAG) - /projects/blast - B.L.A.S.T. (Document Intelligence) - /projects/marketplace-monitor - Marketplace Monitoring & Alerting (Automation & Monitoring) - /projects/menu-pipeline - Menu Data Pipeline (Automation & Monitoring) - /projects/nightshade - Nightshade (Engineering Experiments) - /projects/quantum-singularity - Quantum Singularity (Engineering Experiments) - /blog - Technical Notes, Benchmarks & Architectural Teardowns - /blog/building-grounded-rag-system-without-hallucinations - Building Grounded RAG Systems: Eliminating Hallucinations in Production (AI & RAG) - /blog/fail-closed-web-scraping-sqlite-wal-automation - Fail-Closed Web Scraping: Preventing Silent Data Loss in Python Automation (Automation & Monitoring) - /blog/nextjs-15-production-architecture-patterns - Next.js 15 App Router: 7 Production Architecture Patterns for AI Apps (Full-Stack Web) - /blog/rapidocr-vs-tesseract-document-intelligence-benchmark - RapidOCR vs Tesseract: A Benchmark on Multi-Column PDF Layouts (Document Intelligence) - /uses - Complete Hardware, Software & AI Tooling Stack - /contact - Direct Contact & Inquiry Form - /assistant - Live Interactive AI Portfolio Assistant (Ibrahim GPT) - /pricing.md - Rate Card & Engagement Model Explainer - /okf/index.md - Open Knowledge Format Bundle for AI Agents - /llms.txt - Standardized Machine-Readable Summary - /llms-full.txt - Complete Unabridged Knowledge Base & Technical Context - /api/v1/mcp - Model Context Protocol (MCP Streamable JSON-RPC 2.0 endpoint) - /api/v1/openapi.json - OpenAPI 3.1 REST API Specification - /api/v1/projects - Programmatic Projects JSON Endpoint - /api/v1/services - Engineering Services JSON Endpoint - /api/v1/availability - Real-time Freelance Availability & SLA Endpoint - /api/v1/hire-quote - Automated Project Scope & Quote Calculator API - /.well-known/mcp.json - Remote MCP Server Discovery Manifest - /.well-known/agent-card.json - A2A Agent Protocol Manifest - /.well-known/ai-plugin.json - OpenAI Plugin / Custom GPT Manifest - /.well-known/security.txt - RFC 9116 Vulnerability Disclosure Policy - /feed.xml - RSS 2.0 Syndication Feed - /feed.json - JSON Feed 1.1 Syndication Feed - /humans.txt - Developer Colophon & Credits ======================================== # Complete Pricing & Engagement Guide ======================================== # Pricing & Engagement Guide - Ibrahim Salman (Freelance Full-Stack & AI Engineer) ## Pricing Philosophy Transparent, scope-based pricing with zero hidden agency markups. Because RAG architectures, OCR pipelines, and automation backends vary significantly in complexity and data scale, projects are quoted based on defined scope and clear milestones rather than arbitrary hourly billing. ## Engagement Tiers 1. **Fixed-Scope Deliverables** (Most Popular) - Ideal for: Turnkey AI assistants (RAG), custom OCR/document intelligence pipelines, marketplace scrapers with alerting, or full-stack Next.js web applications. - Structure: 50% on kickoff, 50% upon verified deployment and handover. - Deliverables: Production-ready codebase, complete automated test suite (Pytest/Vitest/Playwright), infrastructure deployment (Vercel/Cloud), and documentation. 2. **Sprint-Based / Ongoing Retainer** - Ideal for: Fast-moving teams needing dedicated engineering capacity for ongoing feature development, pipeline scaling, model fine-tuning, or system maintenance. - Structure: Billed bi-weekly or monthly per sprint with agreed deliverables. 3. **Architecture Audits & Technical Consulting** - Ideal for: Teams needing code reviews, failure mode analysis, RAG retrieval optimization, OCR benchmark audits, or anti-scraping evaluations. ## What Determines the Quote - **Scope & Functional Requirements** - Defined features, integrations, and success criteria. - **Timeline & Urgency** - Standard sprint velocity vs. accelerated critical deadlines. - **Data & System Complexity** - Single-endpoint utility vs. multi-source concurrent pipeline with auth, vector embeddings, multi-LLM fallbacks, and rate limiting. ## How to Get a Quote Email **ibrahim.pk848@gmail.com**, submit the scoping form at **https://ibrahimsalman.vercel.app/hire**, or use **https://ibrahimsalman.vercel.app/contact**. - **Response time**: Within 24 hours. - **Location**: Islamabad, Pakistan (UTC+5 / PKT) - works remotely with clients across North America, Europe, the Middle East, and Asia. ## Verified Production Proof (Evidence of Shipping) - **UET GPT** - Production RAG assistant with multi-provider LLM routing at https://uet-gpt.vercel.app (sign-in required to use it) - **B.L.A.S.T.** - OCR engine, public demo, 664-test suite, 61.6% CER reduction measured on a 14-page benchmark - **Marketplace Monitoring & Alerting** - Fail-closed monitoring system with 299 tests across 4 Python versions × 2 OSes in CI Full portfolio: https://ibrahimsalman.vercel.app/projects ## Current Availability Available for new freelance projects, contract engagements, and consulting roles. --- Last updated: 2026-08-30 ======================================== # Complete Open Knowledge Format Bundle ======================================== ======================================== # Document: /okf/index.md ======================================== --- type: Index title: Ibrahim Salman - Open Knowledge Format Bundle description: Cross-linked concept files describing Ibrahim Salman, his services, hiring guide, and shipped projects, for agent consumption. resource: https://ibrahimsalman.vercel.app/okf/index.md timestamp: 2026-08-30 --- # Ibrahim Salman - Knowledge Bundle An agent-readable index of this site's content, per the [Open Knowledge Format](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/HEAD/okf/SPEC.md) spec. ## Person & Hiring - [Ibrahim Salman](ibrahim-salman.md) - who he is, what he does, availability - [How to Hire Ibrahim Salman](hire-me.md) - engagement models, why hire him, timeline & pricing ## Services - [Services overview](services.md) - the four things he's hired for ## Projects - [UET GPT](uet-gpt.md) — live RAG assistant for university students - [B.L.A.S.T.](blast.md) — OCR automation engine - [Marketplace Monitoring & Alerting](marketplace-monitor.md) — reliability-first monitoring system - [Menu Data Pipeline](menu-pipeline.md) — concurrent scraping + normalization pipeline - [Nightshade](nightshade.md) — source code protection against AI scraping - [Quantum Singularity](quantum-singularity.md) — WebGL2 GPU particle simulation ## Technical Notes & Blog - [Technical Blog](https://ibrahimsalman.vercel.app/blog) - empirical engineering breakdowns and benchmarks - [Grounded RAG Architecture](https://ibrahimsalman.vercel.app/blog/building-grounded-rag-system-without-hallucinations) - multi-provider LLM routing & negative-constraint prompting - [RapidOCR Benchmark](https://ibrahimsalman.vercel.app/blog/rapidocr-vs-tesseract-document-intelligence-benchmark) - 61.6% CER reduction on multi-column PDFs - [Fail-Closed Scraping](https://ibrahimsalman.vercel.app/blog/fail-closed-web-scraping-sqlite-wal-automation) - SQLite WAL concurrency & zero silent loss - [Next.js 15 Patterns](https://ibrahimsalman.vercel.app/blog/nextjs-15-production-architecture-patterns) - React 19 RSC & streaming AI UI ## Reference - [FAQ](faq.md) - common questions about availability, pricing, and the work ======================================== # Document: /okf/hire-me.md ======================================== --- type: Article title: How to Hire Ibrahim Salman - Full-Stack AI & Automation Engineer description: Complete hiring guide, technical capabilities, engagement models, and delivery guarantees for working with Ibrahim Salman. resource: https://ibrahimsalman.vercel.app/hire tags: [hire, freelance, full-stack, ai-engineer, rag, nextjs, python, remote] timestamp: 2026-08-30 --- # How to Hire Ibrahim Salman Ibrahim Salman is an available full-stack developer and AI engineer based in Islamabad, Pakistan (UTC+5), working with clients remotely worldwide. ## Why Hire Ibrahim Salman 1. **Production-Tested AI & RAG**: Builds grounded RAG systems with vector search, multi-provider LLM routing (Groq, Google Gemini, Cerebras), rate limiting, and custom Crawl4AI ingestion pipelines (flagship: [UET GPT](uet-gpt.md)). 2. **Measurable Benchmarked Results**: Engineered [B.L.A.S.T.](blast.md) OCR to deliver a measured 61.6% CER reduction and 3.9x faster CPU latency with 664 automated tests; built [Marketplace Monitor](marketplace-monitor.md) with 299 tests across 4 Python versions and 2 OSes in CI. 3. **Architectural Discipline**: Systems handle rate limits, network retries with exponential backoff, schema validation (Pydantic), and database WAL journaling to eliminate silent failures. 4. **Direct High-Velocity Collaboration**: Direct communication with the engineer building your product - no account managers, no agency overhead, transparent milestone delivery. ## Core Capabilities - **Full-Stack Web**: Next.js 15 (App Router), React 19, TypeScript (strict), Tailwind CSS. - **AI & RAG Systems**: Vercel AI SDK, Groq, Google Gemini, Cerebras, Convex, Vector Embeddings, Crawl4AI. - **Document Intelligence**: Python 3.12, RapidOCR/ONNX, PyMuPDF, SQLite. - **Data Automation**: httpx (HTTP/2), Pydantic, pandas, BeautifulSoup4, Selenium, Playwright, Tenacity. ## Engagement Tiers - **Fixed-Scope Builds**: Defined deliverable (AI assistant, OCR engine, scraper, web app) with clear milestones. - **Sprint-Based Retainers**: Dedicated capacity for ongoing feature delivery and pipeline scaling. - **Technical Consulting**: Architecture audits, RAG evaluation, and benchmark analysis. ## How to Kick Off a Project Email **ibrahim.pk848@gmail.com** or submit the project scoping form at [Hiring Page](https://ibrahimsalman.vercel.app/hire). Response within 24 hours. ======================================== # Document: /okf/ibrahim-salman.md ======================================== --- type: Person title: Ibrahim Salman description: Software engineer specializing in AI automation, monitoring, and document intelligence, available for freelance projects. resource: https://ibrahimsalman.vercel.app/about tags: [freelance, software-engineer, ai, rag, ocr, automation] timestamp: 2026-08-19 --- # Ibrahim Salman Software engineer based in Islamabad, Pakistan (UTC+5), specializing in AI automation, monitoring, and document intelligence - RAG chatbots, OCR pipelines, marketplace/source monitoring with alerting, and workflow automation. Available for freelance projects, works with clients remotely worldwide, replies within 24 hours. ## Contact - Email: ibrahim.pk848@gmail.com - GitHub: https://github.com/Ibrahim-Salman19 - LinkedIn: https://www.linkedin.com/in/ibrahim-salman-dev/ - Upwork: https://www.upwork.com/freelancers/~013e1c54e9a3f7a2b8 ## Education University of Engineering and Technology, Taxila. ## What He's Hired For See [services](services.md) and [how to hire him](hire-me.md). ## Proof of Work See [UET GPT](uet-gpt.md), [B.L.A.S.T.](blast.md), [Marketplace Monitoring & Alerting](marketplace-monitor.md), [Menu Data Pipeline](menu-pipeline.md), [Nightshade](nightshade.md), and [Quantum Singularity](quantum-singularity.md). ## Pricing No fixed rate card - see https://ibrahimsalman.vercel.app/pricing.md ## Common Questions See [FAQ](faq.md). ======================================== # Document: /okf/services.md ======================================== --- type: Article title: Services - Ibrahim Salman description: The four service areas Ibrahim Salman is hired for, each backed by a live shipped project as proof. resource: https://ibrahimsalman.vercel.app/ tags: [services, rag, ocr, monitoring, automation] timestamp: 2026-08-30 --- # Services Four areas, each with a live project as proof rather than a claim. For engagement details and process, see [how to hire Ibrahim Salman](hire-me.md). ## AI Assistants & RAG Turn your docs or site into a grounded AI assistant. Proof: [UET GPT](uet-gpt.md). ## Document Intelligence Extract and structure PDFs, scans, and images. Proof: [B.L.A.S.T.](blast.md). ## Monitoring & Extraction Track changing sources, deliver structured alerts. Proof: [Marketplace Monitoring & Alerting](marketplace-monitor.md). ## Workflow Automation Connect sites, APIs, databases, and notifications. Proof: [Menu Data Pipeline](menu-pipeline.md). ## Pricing & Hiring See https://ibrahimsalman.vercel.app/pricing.md and [how to hire Ibrahim](hire-me.md). ======================================== # Document: /okf/uet-gpt.md ======================================== --- type: Project title: UET GPT description: A ChatGPT-style AI assistant for University of Engineering and Technology students, answering questions with retrieval-augmented generation over real university data. resource: https://ibrahimsalman.vercel.app/projects/uet-gpt tags: [rag, llm, nextjs, convex, live] timestamp: 2026-08-19 --- # UET GPT Status: **Live**. A ChatGPT-style assistant that answers UET students' questions (admissions, fees, courses) using real university data instead of guessing. ## Stack Next.js, TypeScript, Convex (Agent/RAG/Workflow), Vercel AI SDK, Clerk auth, multi-provider LLM routing (Groq, Google Gemini, Cerebras), Upstash rate limiting, Python Crawl4AI ingestion pipeline. ## Links - Live: https://uet-gpt.vercel.app (sign-in required to use it) - Repo: https://github.com/Ibrahim-Salman19/uet-gpt - Case study: https://ibrahimsalman.vercel.app/projects/uet-gpt Proof point for the [AI Assistants & RAG service](services.md). Built by [Ibrahim Salman](ibrahim-salman.md). ======================================== # Document: /okf/blast.md ======================================== --- type: Project title: B.L.A.S.T. OCR Engine description: An OCR automation engine that reads text out of PDFs, slide decks, and images into clean Markdown/Word files. resource: https://ibrahimsalman.vercel.app/projects/blast tags: [ocr, python, document-intelligence, live] timestamp: 2026-08-19 --- # B.L.A.S.T. OCR Engine Status: **Live**. Drop in a PDF, PowerPoint, or image and it extracts the text into clean Markdown/Word files. Measured 61.6% CER (character error rate) reduction and 3.9x faster CPU latency on a 14-page benchmark corpus. ## Stack Python, RapidOCR/ONNX, Streamlit, SQLite. 664-test suite. ## Links - Live demo: https://ocr-book.streamlit.app/ - Repo: https://github.com/Ibrahim-Salman19/OCR - Case study: https://ibrahimsalman.vercel.app/projects/blast Proof point for the [Document Intelligence service](services.md). Built by [Ibrahim Salman](ibrahim-salman.md). ======================================== # Document: /okf/marketplace-monitor.md ======================================== --- type: Project title: Marketplace Monitoring & Alerting description: A reliability-first system that watches job and freelance marketplaces for new postings and emails a structured digest. resource: https://ibrahimsalman.vercel.app/projects/marketplace-monitor tags: [monitoring, python, automation, reliability] timestamp: 2026-08-19 --- # Marketplace Monitoring & Alerting Status: **Complete**. Watches job and freelance marketplaces for new postings and sends a structured email digest the moment one appears. Fail-closed on every external dependency. ## Stack Python, Selenium, SQLite (WAL-journaled), SMTP, Pytest. 299 tests including fuzz and adversarial suites, run in CI across 4 Python versions x 2 operating systems. ## Note Repo kept private for platform trust & safety reasons. ## Links - Case study: https://ibrahimsalman.vercel.app/projects/marketplace-monitor Proof point for the [Monitoring & Extraction service](services.md). Built by [Ibrahim Salman](ibrahim-salman.md). ======================================== # Document: /okf/menu-pipeline.md ======================================== --- type: Project title: Menu Data Pipeline description: A concurrent Python pipeline that scrapes menu data from 17 restaurant sites, validates and normalizes it, and pushes results into a live Google Sheet. resource: https://ibrahimsalman.vercel.app/projects/menu-pipeline tags: [automation, python, web-scraping, data-pipeline] timestamp: 2026-08-19 --- # Menu Data Pipeline Status: **Complete**. Concurrently pulls menu data from 17 restaurant sites, validates and cleans it, and pushes a unified dataset straight into a live Google Sheet. ## Stack Python, httpx (HTTP/2), Pydantic, pandas, Google Sheets API, Tenacity. ## Links - Repo: https://github.com/Ibrahim-Salman19/foodmenu_scraper - Case study: https://ibrahimsalman.vercel.app/projects/menu-pipeline Proof point for the [Workflow Automation service](services.md). Built by [Ibrahim Salman](ibrahim-salman.md). ======================================== # Document: /okf/nightshade.md ======================================== --- type: Project title: Nightshade description: A Java 21 tool that poisons source code against unauthorized AI training scrapes by rewriting it to stay functional for humans and compilers while degrading as training data. It does not prevent the scrape itself. resource: https://ibrahimsalman.vercel.app/projects/nightshade tags: [java, ast, ai-training-poisoning, security] timestamp: 2026-08-19 --- # Nightshade Status: **Released, v3.5.0**. Applies eight obfuscation strategies over an AST, validated through compilation and regression testing, so code keeps working for humans and compilers while degrading as training data. Ships as a CLI, GitHub Action, and pre-commit hook. ## Stack Java 21, Maven, AST-based transformation, GitHub Actions, JUnit 5. ## Links - Repo: https://github.com/Ibrahim-Salman19/night-shade - Case study: https://ibrahimsalman.vercel.app/projects/nightshade Built by [Ibrahim Salman](ibrahim-salman.md). ======================================== # Document: /okf/quantum-singularity.md ======================================== --- type: Project title: Quantum Singularity description: A real-time WebGL2 particle simulation with up to 110,000 GPU particles, controlled by hand gestures and live audio, running entirely in-browser from a single static HTML file. resource: https://ibrahimsalman.vercel.app/projects/quantum-singularity tags: [webgl2, three-js, creative-coding, live] timestamp: 2026-08-19 --- # Quantum Singularity Status: **Live**. A real-time WebGL2 particle simulation (up to 110,000 GPU particles) driven by hand gestures and live audio, running entirely on-device from a single static HTML file. ## Stack Three.js, WebGL2, MediaPipe (hand tracking), Web Audio API, Playwright (testing). ## Links - Live: https://animation-zeta-rosy.vercel.app/ - Repo: https://github.com/Ibrahim-Salman19/quantum-singularity - Case study: https://ibrahimsalman.vercel.app/projects/quantum-singularity Built by [Ibrahim Salman](ibrahim-salman.md). ======================================== # Document: /okf/faq.md ======================================== --- type: FAQ title: FAQ - Ibrahim Salman description: Common questions about hiring, availability, technical capabilities, pricing, and project proofs. resource: https://ibrahimsalman.vercel.app/ tags: [faq, hire, freelance, full-stack, ai] timestamp: 2026-08-30 --- # FAQ ## Is Ibrahim Salman available for freelance projects and hiring? Yes. Currently available, replies within 24 hours. Based in Islamabad, Pakistan (UTC+5), works with clients remotely worldwide across US, Europe, and Asia. See [how to hire him](hire-me.md). ## Why hire Ibrahim Salman over an agency or generalist developer? Direct engineer communication with zero agency overhead. Ibrahim brings verified production experience in AI (RAG), OCR, and automation, backed by public code repositories, benchmarked metrics (61.6% CER reduction, 3.9x speedup in [B.L.A.S.T.](blast.md)), and rigorous test suites (299+ and 664+ tests in CI). ## What kind of projects does he build? Four areas: AI assistants and RAG apps, document intelligence and OCR, monitoring pipelines that watch changing sources and alert on new activity, and workflow automation connecting sites, APIs, and databases. See [services](services.md). ## What technologies does he use? Frontend: Next.js 15, React 19, TypeScript, Tailwind CSS. Backend & Data: Python 3.12, FastAPI, Convex, SQLite, PostgreSQL, Pandas, Pydantic. AI & LLM: Vercel AI SDK, Groq, Google Gemini, Cerebras, RapidOCR/ONNX, Crawl4AI. ## How does he prevent AI hallucinations and handle errors? By designing strict retrieval-augmented generation (RAG) architectures with multi-provider LLM fallbacks, where answers are strictly grounded in vector-retrieved university or company data, admitting uncertainty when context is missing. ## How do I start a project? Email ibrahim.pk848@gmail.com with what you're building and what's broken. He replies within 24 hours. ## How is pricing determined? Pricing is milestone or scope-based, quoted after an initial requirements conversation - no hidden fees or agency markups. Full breakdown: https://ibrahimsalman.vercel.app/pricing.md ## Are the portfolio projects real, working software? Yes. [UET GPT](uet-gpt.md) is live in production (sign-in required to use it), [B.L.A.S.T.](blast.md) has a public demo backed by a 664-test suite, and [Marketplace Monitoring & Alerting](marketplace-monitor.md) shipped with 299 tests (including fuzz/adversarial suites) across 4 Python versions and 2 OSes in CI.