Mubarak Muse
Verified Expert in Engineering
Amazon Web Services (AWS) Developer
Minneapolis, MN, United States
Toptal member since March 22, 2022
Mubarak is a software engineer with 8 years of experience building AI systems and production back ends. As a senior software engineer at Brighter Tunnel, an AI and software consulting firm, he ships LLM-backed agents end-to-end—most recently a healthcare AI platform that automates insurance prior authorizations. He previously built real-time freight systems at Amazon and healthcare benefits infrastructure at Collective Health. He works in Python, FastAPI, React, and TypeScript, deploying on AWS.
Portfolio
Experience
- Full-stack - 8 years
- Amazon Web Services (AWS) - 8 years
- Python - 8 years
- TypeScript - 7 years
- FastAPI - 7 years
- Artificial Intelligence (AI) - 5 years
- Claude - 3 years
Work Experience
Senior Software Engineer
Brighter Tunnel
- Built and shipped an AI and rules engine that reads a clinical note, extracts diagnoses and procedures, and drafts a grounded prior authorization letter for payer submission.
- Architected Opus Eternal, an agentic multi-stage pipeline that transcribes interview audio and orchestrates LLMs to produce edited book drafts, using Python, FastAPI, and LangGraph.
- Built a multi-tenant AI-native learning management platform with AI-assisted lesson planning, grading, and summarization on Next.js, TypeScript, Supabase, and OpenAI.
- Delivered an EV charging platform that extracted real-time station telemetry over the OCPP protocol into multi-tenant operator dashboards.
Interview Engineer
Karat
- Conducted 1,000+ technical interviews for companies including PayPal, Palantir, The New York Times, and Lyft.
- Evaluated candidates on data structures, algorithms, and system design against calibrated rubrics, writing assessments that fed directly into hiring decisions.
- Proposed and evaluated interview content, helping refine the questions and rubrics used across the screening process.
Software Engineer
Amazon.com
- Built an event-driven Python service that reroutes trucks to alternate terminals in real time when a terminal goes down. AWS Step Functions, Lambda, SQS with dead-letter queues.
- Implemented dead-letter queues and retry handling so that freight rerouting events survived downstream failures during terminal outages.
- Built a full-stack app capturing real-time terminal data (shipments, pallets, problem-solving cases) for live operational visibility. Used DynamoDB, React, TypeScript, Java, Kotlin, and AWS CDK.
Software Engineer
Collective Health
- Led the plans and benefits domain during a monolith-to microservices migration; built two Java Spring Boot services (120+ API endpoints) and a Node.js parsing layer integrating GraphQL, GitHub, and Airtable APIs.
- Designed flexible, highly configurable data models for custom plan structures beyond standard benefit design; the source of truth for ~13,500 plans across 100+ clients (Uber, Pinterest, eBay), consumed by claims, eligibility, and data science teams.
- Built Python automation on a Kubernetes CronJob that detected ICD/CPT code changes and auto-updated the source of truth, replacing a manual multi-API workflow; added a PostgreSQL audit trail with Datadog and Prometheus monitoring.
- Delivered full-stack features (React, TypeScript, Angular) on PostgreSQL-backed services; deployed via Docker, Kubernetes, and Jenkins CI/CD.
Project History
Prior Auth AI Agent
The system reads a clinical note, extracts diagnoses and procedures, and matches them to CPT and ICD-10 codes. The matching layer is deliberately deterministic rather than LLM-inferred because a hallucinated procedure code results in a denied claim and a compliance problem. The LLM handles what it is actually good at: reading unstructured clinical language and drafting a grounded letter that cites the specific policy criteria the case satisfies.
I designed and built the entire system: a FastAPI back end, a React front end, Claude for extraction and drafting, and AWS infrastructure using ECS, SQS, DynamoDB, and CloudFront. Infrastructure is defined in Terraform with CI/CD through GitHub Actions.
Opus Eternal: Agentic Audio-to-book Pipeline
I architected a multi-stage agentic pipeline that handles the full path. A transcription stage converts audio to text. A structuring stage identifies narrative threads, chronology, and recurring themes across sessions that were never recorded in order. The drafting stage produces chapter text in the speaker's own voice, and the editing stage revises it for continuity and pacing.
The interesting engineering problem was orchestration rather than generation. Each stage needs a different context, different failure handling, and a way for a human editor to intervene without restarting the pipeline. I built it in Python and FastAPI with LangGraph managing stage transitions, and a React interface for the human-in-the-loop review steps.
Capture Flow — Turn Expert Browser Workflows into AI Playbooks
https://github.com/mubarakMuse/capture-flowInbox Concierge — AI Gmail Triage into Priority Buckets
Education
Bachelor's Degree in Computer Science
University of Minnesota - Minneapolis, Minnesota, USA
Skills
Libraries/APIs
Node.js, React, REST APIs, Stripe, Pydantic, OpenAI API, Claude API, OpenAPI, Gmail API
Tools
ChatGPT, GitHub, Zoom, Claude Code, Apache Airflow, Sequelize, Claude, AWS Step Functions, Terraform, Amazon CloudFront, Amazon Simple Queue Service (SQS)
Languages
Java, Python, JavaScript, TypeScript, SQL, CSS, HTML, Go, GraphQL, Kotlin
Frameworks
Next.js, Agentic Frameworks, Hibernate, Spring Boot, Spring, Express.js, JSON Web Tokens (JWT), Jest, Tailwind CSS, Django, NestJS, LangGraph
Paradigms
Microservices, Event-driven Architecture, Real-time Systems, Automation, Back-end Architecture, Role-based Access Control (RBAC), Agile, Scrum, DevOps, Marketplace Platforms, HIPAA Compliance, Model Context Protocol (MCP), Database Design
Platforms
Java EE (Jakarta EE), WordPress, Vercel, Amazon Web Services (AWS), Docker, Kubernetes, Apache Kafka, Google Cloud Platform (GCP), AWS Lambda
Storage
PostgreSQL, Databases, MySQL, Redis, Relational Databases, MongoDB, Database Management, NoSQL, Amazon DynamoDB, Amazon S3 (AWS S3)
Industry Expertise
Healthcare
Other
Software Engineering, Cursor AI, Full-stack, Full-stack Development, APIs, Interviewing, Communication, Screeners, Team Leadership, Software Architecture, Architecture, Agentic AI, CI/CD Pipelines, Infrastructure, AI-generated Code, Artificial Intelligence (AI), Web Development, Healthtech, Workflow, Workflow Automation, Integrations, Sports, Debugging, Workflow Automation & System Integration, Minimum Viable Product (MVP), Figma to React, Multi-agent Orchestration, AI Integration, Healthcare Software, Technical Architecture, Education Technology (Edtech), Training, Concurrency, Event-driven Systems, Agentic Workflow Design, Agentic AI Systems, Anthropic, Back-end, Product Mindset, Startups, AI Tools, Amazon RDS, Airtable, OpenAI, FastAPI, Data Analytics, Data Engineering, API Integration, Authentication, Cloud Storage, Supabase, Supabase Auth, Code Review, Large Language Models (LLMs), AI-assisted Development, Cloudflare, Infrastructure as Code (IaC), Observability, Orchestration, Fintech, OAuth, Cloud Architecture, Compliance, Data Security, HIPAA, Product Management, SaaS, CTO, Bun, Reviews, Security, RAG Systems, Pinecone, Marketplaces, Retrieval-augmented Generation (RAG), Vector Databases, Electronic Health Records (EHR), Distributed Systems, Optical Character Recognition (OCR), State Machines, Multi-tenant SaaS, AI Agents, Agentic RAG Systems, Multi-tenant Architecture, SOC 2, Financial APIs, Front-end, Logistics, Order Management, Shipping, Product Development, Machine Learning, Research, Analytics, LangChain, GitHub Actions, Speech-to-Text (STT), Data Visualization, Charts, Chrome Extensions, AWS App Runner
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring