I build production ML & LLM systems — deployment, monitoring, and real-world AI infrastructure.

Currently at iO Health. Open to freelance ML, LLM, and MLOps projects.

📧 baraalsedih@gmail.com · LinkedIn · Portfolio

A modular collection of ML, MLOps, and LLM systems focused on production-style engineering, system design, and reusable AI infrastructure.

All of my work are in this org: Atlas AI

MLOps + LLMOps observability system for monitoring ML and LLM systems in production-style environments.

• Model performance monitoring
• Data drift detection and alerting design
• LLM monitoring for RAG and prompt-based systems (quality, latency, tokens)
• Incident routing (auto vs manual escalation)

Retrieval-Augmented Generation system built over a personal knowledge base.

• Document ingestion and chunking pipeline
• Vector-based retrieval system
• Context-aware response generation

ML system design for account takeover (ATO) / anomaly detection.

• Fraud detection pipeline design
• Feature engineering and modeling workflow
• Production-style ML architecture

CLI tool that starts all components of an ML/LLM application (API, model service, and dependencies) and blocks until everything is running and healthy.

• Spins up backend API, model services, and required infrastructure
• Manages startup order and dependencies automatically
• Waits for health checks to confirm every service is running
• Prints a “ready” state when the full system is operational

Python + FastAPI + Docker + MLflow + AWS + LLM APIs

• Production ML & LLM systems (deployment, scaling, reliability)
• MLOps & observability systems (monitoring, drift, alerting)
• LLM applications (RAG, pipelines, AI workflows)
• Backend infrastructure for AI systems (FastAPI, Docker, AWS)