Hi! I'm Oussama Sallak, a full-stack and AI engineer specializing in building intelligent systems with AI agents, LangChain, and MLOps pipelines. I develop scalable AI-powered applications from model deployment to production monitoring, combining strong backend expertise with modern AI integration. Currently leading AI integration at Chemp.ai and building MLOps infrastructure at Andala.ai.

• AI & Machine Learning
  • LangChain, RAG (Retrieval-Augmented Generation), PyTorch
  • Vector Databases, Agentic Workflows
  • Model Deployment, Fine-tuning, Inference Optimization
• Backend & MLOps
  • Python, FastAPI, Django
  • Docker, Kubernetes, AWS, CI/CD
  • PostgreSQL, Redis, Celery, RabbitMQ
• Frontend & Mobile
  • React, Next.js, React Native, TypeScript
  • TailwindCSS, Framer Motion
• Systems Programming
  • C, C++ (UNIX, Multithreading, Networking)
  • Performance Optimization, Low-level Programming

End-to-end MLOps infrastructure for deploying multimodal models (music, video, image generation). Built scalable inference APIs with PyTorch and Django, implemented CI/CD pipelines for automated model deployment, and reduced inference costs by 45% through GPU optimization and caching strategies.

• Tech Stack: Python, Django, PyTorch, AWS, Docker, CI/CD

AI-powered e-learning platform with LangChain-based retrieval-augmented generation (RAG). Integrated domain-specific LLMs, vector databases, and personalized tutoring agents enabling real-time adaptive learning with sub-second retrieval latency for hundreds of concurrent users.

• Tech Stack: Python, FastAPI, LangChain, RAG, Redis, Vector DBs, Docker

Real-time multiplayer pong game with WebSocket communication and microservices architecture. Supports hundreds of simultaneous players with minimal latency, featuring chat, matchmaking, authentication, and comprehensive end-to-end testing.

• Tech Stack: NestJS, TypeScript, WebSocket, PostgreSQL, Docker

Netflix-style streaming platform with dynamic content loading, server-side rendering, and optimized caching. Implemented analytics and user behavior tracking for personalized recommendations, reducing page load time by 40%.

• Tech Stack: Next.js, TypeScript, PostgreSQL, TailwindCSS

RFC 1459-compliant IRC server with efficient concurrency and event handling. Implemented message routing, user authentication, and command parsing demonstrating mastery of sockets, multi-threading, and low-level networking.

• Tech Stack: C++, Network Programming, Unix, Concurrency

• 1337 School (42 Network) - Completed Common Core at 100%, currently pursuing Senior Software Architect specialization with focus on AI/MLOps and distributed systems
• Advanced Certifications - Python, JavaScript, C/C++ (CodinGame)

• Leading AI integration efforts at Chemp.ai (mobile & web platforms)
• Building production MLOps systems at Andala.ai
• Developing AI-assisted learning platforms with RAG and agentic workflows
• Exploring advanced retrieval techniques and multimodal model deployment

• Portfolio: osallak.tech
• GitHub: github.com/osallak
• LinkedIn: linkedin.com/in/osallak
• Email: oussamasallak1@gmail.com

I'm passionate about building intelligent systems that are scalable, reliable, and cost-efficient. If you're working on AI/ML projects or want to collaborate, let's connect! 🚀