AI Engineer | RAG Pipelines, Agentic AI, FastAPI Backends & Computer Vision Solutions

I am an AI Engineer specializing in building production-grade AI systems, secure backend architectures, and end-to-end intelligent applications. I focus on delivering scalable, verifiable, and explainable AI products with clean code and a security-first design.

🛠️ Core Capabilities
Generative AI & Agentic AI: LLaMA 3.3, LangChain, LlamaIndex, LLM Orchestration, Prompt Engineering & Hallucination Guardrails.

Retrieval-Augmented Generation (RAG): Production RAG pipelines, ChromaDB vector indexing, semantic chunking, and BGE-Large embeddings.

Backend & API Engineering: FastAPI, AsyncIO, SQLAlchemy Async, PostgreSQL (asyncpg), Redis, Pydantic v2, JWT dual-token auth, and REST APIs.

Computer Vision & ML: TensorFlow, PyTorch, CNNs, Transfer Learning, MobileNetV3, Scikit-learn, and real-time inference engines.

DevOps & Architecture: Docker, CI/CD with GitHub Actions, clean architecture, and automated testing (pytest).

🚀 Featured Project Highlights
YordamAI (Citizen Services Assistant): Solo-built a multilingual RAG-powered chatbot (Uzbek, Russian, English) serving 300+ knowledge base chunks using LLaMA 3.3 70B, LangChain, ChromaDB, and Telegram API.

Production RAG Chatbot Pipeline: Built an end-to-end RAG system with HNSW vector retrieval, strict hallucination guardrails, and source-metadata verification.

Teachable Machine AI Platform: Developed a full-stack, decoupled FastAPI + Docker ML platform supporting real-time webcam inference and multi-architecture model training (89% accuracy).

Async Security Microservice: Architected a high-throughput authentication backend using FastAPI, dual-token JWTs, Redis token revocation, and PostgreSQL connection pooling.

Whether you need a full-stack AI platform, custom RAG chatbot, custom agentic workflows, or a secure FastAPI backend, I deliver clean, production-ready code with complete transparency and automated tests.

Let’s discuss your project requirements!

Work Terms

• Payment & Milestones: Prefer Fixed-Price project milestones tied to clear deliverables, or Hourly contracts via Guru SafePay for ongoing development.
• Communication: Available via Guru Messages for primary project discussions and documentation. Video/audio calls (Zoom/Google Meet) can be arranged for initial kickoffs and technical alignments.
• Time Zone & Hours: Operating in PKT (UTC+5), with flexible overlap for client working hours across US, European, and Asian time zones.
• Code Delivery & Quality: All code is delivered clean, modular, and fully tested (pytest) with complete documentation, Git repositories, and post-deployment support.