Yordanos Abat

About

Senior AI Full-Stack Engineer specializing in production LLM systems, RAG pipelines, and agentic AI workflows. Builds scalable AI products end-to-end using Python, FastAPI, LangChain/LangGraph, vector databases, React, and modern cloud infrastructure, with experience delivering healthcare AI solutions including clinical documentation copilots, transcription pipelines, and HIPAA-compliant EHR integrations. Previously built large-scale full-stack and platform systems for Netflix Ads and enterprise digital platforms for Fortune 500 clients at Accenture.

Experience

• Senior AI Full-Stack Engineer · Ambience Healthcare08-01-2022

  • Delivered Ambience Clinical Copilot, an AI-powered documentation platform used during patient encounters, implementing a production RAG architecture using Python, FastAPI, LangChain, pgvector, and OpenAI models to generate structured clinical notes integrated with major EHR systems. • Built a real-time clinical transcription pipeline using OpenAI Whisper, asynchronous FastAPI services, and streaming APIs that converted doctor-patient conversations into structured SOAP drafts during visits, reducing manual documentation time for clinicians. • Designed scalable retrieval pipelines for medical knowledge grounding, implementing embedding generation, chunking strategies, and hybrid vector search using PostgreSQL pgvector and Redis caching to deliver sub-second retrieval of historical patient context. • Built agentic prompt workflows using LangGraph to orchestrate multi-step reasoning across transcripts, diagnoses, and patient history, improving note accuracy and reducing hallucination rates in generative outputs. • Implemented token streaming and server-sent events from LLM inference services to React clients, enabling real-time UI updates and significantly improving perceived latency during live chart creation. • Developed robust LLM evaluation and prompt iteration pipelines, implementing structured outputs, prompt templates, and automated regression checks to maintain clinical quality across model upgrades. • Integrated healthcare EHR APIs using Python microservices and secure REST interfaces, enabling bidirectional synchronization of patient data while meeting HIPAA compliance standards. • Built semantic search and contextual retrieval services leveraging embedding models and vector indexing to surface relevant prior encounters and diagnoses across millions of records. • Owned AI feature delivery end-to-end from ideation to production deployment, collaborating with clinicians, product managers, and platform engineers to continuously improve AI-driven clinical workflows. • Practiced AI-assisted development and vibe coding using modern coding copilots and LLM tooling to rapidly prototype AI features, iterate prompt workflows, and accelerate delivery of production-ready AI capabilities.

• Senior Full-Stack Engineer · Netflix01-01-2019 – 06-30-2022

  • Delivered core components of the Netflix Ads Platform Web, building large-scale ad decisioning and targeting systems using React, GraphQL, Java Spring Boot, Kafka streams, and TensorFlow ranking models that powered personalized advertising experiences across streaming sessions. • Designed a low-latency ad decision service using Spring Boot and Kafka event streams to select personalized ads within strict playback constraints, improving ad fill rate and engagement across millions of daily viewers. • Implemented dynamic ad insertion workflows within the React streaming player using GraphQL APIs and server-stitched manifests, enabling seamless transitions between video content and advertisements without increasing rebuffer rates. • Developed telemetry pipelines using React instrumentation and Node.js event collectors to capture high-fidelity engagement signals that powered downstream machine learning training datasets. • Built experimentation hooks integrated with Netflix experimentation frameworks to enable controlled rollout of ad load strategies, frequency capping, and revenue optimization models. • Reduced service latency under peak traffic by refactoring synchronous service dependencies into asynchronous Kafka-driven workflows, maintaining sub-second decision times across distributed services.

• Full-Stack Engineer · Accenture06-01-2016 – 12-31-2018

  • Delivered enterprise digital platforms for Fortune 500 clients using Angular, React, Java Spring Boot, and REST microservices, enabling large-scale modernization initiatives that reduced manual business processing by more than 35%. • Designed backend service architectures using Java, Spring Boot, and Hibernate integrated with Oracle and MySQL, supporting high-volume transactional systems processing hundreds of thousands of daily requests. • Built responsive single page applications using Angular and TypeScript that improved onboarding conversion and reduced customer drop-off across financial services platforms. • Implemented enterprise authentication frameworks using OAuth 2.0, JWT, and SSO integrations to strengthen security posture across regulated environments. • Optimized backend APIs by refactoring service layers and introducing Redis caching strategies, reducing average request latency by more than 40 percent under peak load conditions. • Partnered with client stakeholders, product teams, and architects to translate complex business requirements into scalable technical solutions delivered within aggressive consulting timelines.

Education

• University of CaliforniaMaster's degree, Computer Science

Similar talent on Pangea

EErese R.Senior Full-Stack & AI Engineer | 10 Years Experience (AI/LLMs, RAG, Full-Stack)Hong KongTTravis B.Senior Software Engineer | Full-Stack & Cloud | AI/ML-Enabled PlatformsPhoenix, AZJunaid A.Senior Full-Stack Engineer | 10+ Years | AI Training & LLM IntegrationLahore, PakistanJJamal M.Senior Software and AI Engineer | LLM Systems, RAG, Autonomous AgentsDetroit, MI