AI assistant for a clinical records system
Healthcare. In production.
A RAG assistant inside a patient records system that answers physicians’ questions with cited sources, and says so when it can’t answer.
- RAG
- Guardrails
- LLM evaluation
- React
Fahad Mushtaq, Lahore, Pakistan
I’m a full-stack AI engineer. I build RAG assistants, LLM guardrails and the React and Node.js products around them.
10+ years in software engineering. I work end to end, from AI pipelines and REST APIs to production React and Next.js apps.
Work
Three systems I’ve built, described by what they do. Client systems stay confidential, so the healthcare work is shown as an architecture diagram.
Road defect detection with human review
Infrastructure inspection. Contract.
Computer-vision models detect and classify pavement defects and pre-label the images. Engineers review, correct and approve the labels, which then become training data.
- Computer vision
- Human-in-the-loop
- React
- TypeScript
Subscription commerce automation
Direct-to-consumer telehealth. Contract.
n8n workflows that keep orders, subscriptions and customer data in sync across commerce, subscription and review platforms, plus a subscription-provider migration.
- n8n
- Shopify
- Subscription APIs
About
I’m a Senior AI Full-Stack Engineer with 10+ years of software engineering experience, building production LLM features for healthcare.
I deliver RAG assistants and real-time AI summarization for a value-based care platform used by clinicians and care teams at health systems and ACOs, along with computer-vision data labeling workflows and n8n automation.
I work end to end, from AI pipelines and REST APIs to production React and Next.js web applications.
What I work on
- Generative AI and LLMs
- RAG pipelines, embeddings, semantic search and chunking. Prompt engineering, grounding, summarization and conversational AI. LLM evaluation, monitoring, guardrails and governance.
- Agentic and real-time systems
- Agentic workflow automation with n8n, real-time AI streaming over WebSockets, and enterprise API integration.
- ML data operations
- Computer vision, human-in-the-loop labeling, label QA and dataset curation.
- Full-stack engineering
- React, Next.js and TypeScript. Node.js, Express, MySQL and REST APIs. Reusable component architecture, legacy platform modernization and Shopify.
- Leadership
- Team mentoring, code reviews, engineering standards and cross-functional delivery.
Courses
Agentic AI, Advanced AI, Safe and Reliable AI via Guardrails (DeepLearning.AI)
Education
BS Information Technology, Punjab University College of Information Technology, 2012 – 2016
Experience
2020 – now
Senior AI Full-Stack Engineer, Persivia
USA, remote
- Built a RAG clinical assistant that gives physicians source-cited answers using chunking, embeddings and semantic search.
- Implemented AI safety guardrails: input and output validation, sensitive-data protection, grounded-answer checks and low-confidence fallbacks.
- Set up LLM evaluation and monitoring to track answer relevance, faithfulness and accuracy in production.
- Built a real-time escalation dashboard that uses LLM summarization to surface critical items, SLA risks and owners, with live updates over WebSockets.
- Built a reusable React component library adopted across multiple modules.
- Lead full-stack development of the core platform, including front-end modules, REST API integrations and legacy modernization, and mentor the team through code review.
2018 – 2020
Full-Stack Developer, Cyber Advance Solutions
Lahore
Full-stack web applications end to end: React front ends, Node.js and Express REST APIs and MySQL data layers, from requirements through deployment. Performance work across the stack, from front-end assets to API and database interactions.
2017 – 2018
Front-End Developer, TrangoLabs
Lahore
2016 – 2017
Associate Software Engineer, Wali Systems
Lahore
What people I’ve worked with say
“Fahad's expertise in React and component architecture is impressive. He built a scalable, reusable component library that our entire team adopted — clean code, consistent design system, and zero headaches.”
Sohaib Amjad“Fahad took our outdated frontend and rebuilt it in React from the ground up. The performance improvement was immediate and the codebase was clean, well-structured, and easy for the rest of the team to maintain.”
Andrew YunGet in touch
I’m open to senior AI engineering roles and freelance projects. Email is the fastest way to reach me.