Most AI projects and automation scripts work well in a quick demo, but run into trouble once they hit real-world conditions. APIs timeout, website layouts change, Cloudflare blocks scrapers, or edge cases crash the pipeline.
I help founders and engineering teams build reliable, production-ready AI applications, web automations, and modern full-stack tools that keep running smoothly without constant babysitting.
Over the past few years, I have worked as a core engineer for venture-backed AI companies, including XY.AI Labs building multimodal browser agents and healthcare workflow systems, as well as partnering directly with independent founders on end-to-end products.
I take pride in writing clean, well-tested code, communicating clearly at every step, and leaving codebases in better shape than I found them.
WHAT I CAN HELP YOU BUILD
► AI Agents and Practical Workflow Automation
▪ Long-running, fault-tolerant Python pipelines with automated retries and failure recovery using Temporal, Celery, and Redis.
▪ Human-in-the-loop review stations where AI drafts proposals, claims, data extraction, and reports for human validation before execution.
▪ Controlled, source-grounded LLM systems using OpenAI, Claude, and Gemini with schema validation via Pydantic to avoid hallucinations.
► Browser Automation and Chrome Extensions
▪ Manifest V3 Chrome extensions with background service workers, content scripts, and clean side panel interfaces.
▪ Deep browser automation using Chrome DevTools Protocol, handling shadow DOM, nested iframes, and persistent login states.
▪ Resilient web scraping engines that handle Cloudflare protections, dynamic single-page apps, and pagination without getting stuck.
▪ Model Context Protocol servers and Python SDKs for programmatic browser control.
► Modern Full-Stack Applications
▪ Frontends built with Next.js App Router, React, TypeScript, Tailwind CSS, and shadcn/ui.
▪ Clean, type-safe API layers using tRPC, FastAPI, and gRPC microservices.
▪ Real-time data streaming, interactive charts, and responsive dashboards.
► Data Platforms and Internal Tools
▪ Interactive data apps and internal admin tools built with Streamlit and Next.js.
▪ PostgreSQL database design with SQLAlchemy and clean Alembic migrations.
▪ Governed semantic analytics using Google BigQuery and Cube.dev.
SELECTED PROJECTS FROM MY RECENT WORK
Faisal S. earns an estimated $5.5k/mo. That's 4.8× the typical freelancer and more than 99.62% of everyone we track.
► Multimodal Browser Agent and Chrome Extension for XY.AI: Built a Manifest V3 extension and accompanying Python SDK that records and replays user browser actions via Chrome DevTools Protocol, featuring self-healing DOM diffing and a Model Context Protocol server.
► Enterprise Workflow Orchestration Engine for XY-Mind Platform: Engineered resilient backend services and human-in-the-loop review queues for complex, regulated healthcare operations backed by gRPC and Temporal.
► SkinSmith AI Skincare Analysis SaaS: Built a full-stack SaaS product from scratch. A Next.js, Prisma, and PostgreSQL platform doing multi-model skin photo analysis using Claude and Gemini, free and paid tier gating, Stripe payments, personalized skincare routine generation, and real product recommendations sourced via Serper and GPT-4o structured extraction, alongside a supporting serverless AWS backend.
► Live Bloodstock Analysis Tool for NH Pedigree Scout: Built and deployed a live public app that scrapes UK and Irish auction catalogues while handling Cloudflare, evaluates lots via Google Gemini, and stores pedigree rankings in Supabase PostgreSQL.
► Source-Grounded Brand Intelligence Engine: Designed an AI engine that ingests company knowledge bases and generates validated, brand-safe marketing content with automated claim checking and editorial approval steps.
► Custom Decision Tools and Diagnostic Dashboards: Cancer diagnostic KPI calculation engines, renewable energy use-case scoring tools, and AWS FinOps cost-monitoring dashboards.
MY WORKING STYLE
▪ Clear and Honest Communication: You will always know what was done, what is next, and if any unexpected roadblock comes up.
▪ Surgical Changes: I respect your existing codebase. I do not rewrite working features or introduce unnecessary libraries.
▪ Tested and Documented: Every delivery includes clear setup steps, lockfiles, and test coverage so you or your team can run it with zero friction.
▪ 100 percent Job Success: Every contract I have completed on Upwork has a five-star rating, because I do not consider a milestone done until you are genuinely satisfied.
LET'S DISCUSS YOUR PROJECT
Click the Message or Invite to Job button on the top right, and let us hop on a quick chat to map out your technical requirements.