AI & Machine Learning
PyTorch, TensorFlow, Hugging Face Transformers, Scikit-learn, PyTorch Lightning
Hello, I'm
3+ years of experience in Machine Learning, NLP, Computer Vision, and AI Agent Orchestration. From research to production, from idea to impact.
class AIEngineer:
def __init__(self):
self.name = "Komiljon"
self.role = "AI Developer"
self.skills = [
"LLMs", "RAG",
"Computer Vision",
"Voice AI",
"Multi-Agent Systems"
]
def build(self, idea):
return self.deploy(
self.optimize(
self.train(idea)
)
) # production-readyI am an AI Developer with over 3 years of industry and academic experience in machine learning, NLP, and computer vision.
My expertise spans speech-to-text (STT), large language models (LLMs), retrieval-augmented generation (RAG), AI-generated content detection, and multi-agent orchestration.
I have built and deployed real-time AI systems for voice authentication, DeepFake detection, audio watermarking, and AI-driven text/voice classification, integrating RAG pipelines and agent-based architectures for domain-specific tasks.
I specialize in model quantization and optimization (ONNX, TensorRT, TensorFlow Lite), delivering scalable, efficient AI solutions across mobile and cloud environments.
Multilingual speech recognition, AI security systems, real-time inference acceleration
MSc in Computer Engineering — Chonnam National University, South Korea
Uzbek (Native), English (Professional), Korean & Russian (Working)
PyTorch, TensorFlow, Hugging Face Transformers, Scikit-learn, PyTorch Lightning
LiveKit, Pipecat, Whisper, TTS, VAD, Real-time Voice Pipelines
LangChain, LangGraph, RAG systems, Vector DB (Pinecone, Qdrant, FAISS), Multi-Agent Orchestration
Docker, Kubernetes, MLflow, FastAPI, TorchServe, AWS, GCP, Azure
Selected projects focused on AI security, real-time inference, and advanced computer vision.
AI-based audio spoofing detection for voice authentication systems. Built with RawNet2 and deployed with FastAPI for real-time inference.
GitHubVoice phishing detection system analyzing speech for fraudulent audio. Implemented as a streaming classification pipeline.
GitHubYOLOv5-based license plate detection with ONNX Runtime acceleration for image, video, and webcam input.
GitHubReal-time gesture recognition system for Uzbek sign language, built to improve accessibility and communication.
GitHubI’m available for AI engineering roles, contract work, and collaboration on speech, vision, and production ML systems.