BekhruzNutfilloev

Machine Learning Engineer — building production AI at national scale. Tashkent, Uzbekistan.

View WorkResume ↓
Experience — 5+ yearsMyID requests / month — 0M+Banks served — 0+Biometric accuracy — 99.7%

Selected Work

06 projects
  • 01

    MyID — National Biometric Platform

    End-to-end facial recognition pipeline deployed across 50+ banks nationwide. Processes 10M+ authentication requests monthly at 99.7% accuracy with sub-200ms latency. Liveness detection achieving 98.5% anti-spoofing accuracy.

    PyTorchArcFaceONNXC/C++TensorRTFAISS
    2021–2025Biometrics
    UZINFOCOM
  • 02

    PalmID — Palm Vein Recognition

    Production-ready palm vein recognition with C/C++ SDK for IoT deployment. 60% inference time reduction via ONNX + TensorRT acceleration. AES-256 model encryption on edge devices.

    C/C++ONNX RuntimeTensorRTCUDAIoT
    2022–2024Edge AI
    UZINFOCOM
  • 03

    AI Candidate Matching System

    RAG-based HR platform with semantic search across 1.6M student database. LaBSE multilingual embeddings, LangGraph orchestration, Qdrant vector store. End-to-end query in 3–5 seconds.

    RAGLangGraphQdrantLaBSEFastAPIPostgreSQL
    2025–2026LLM / Search
    AI Impl. Group
  • 04

    Robocontest.uz — Autonomous Robotics

    Computer vision systems for competitive robotics supporting 500+ participants. Real-time object detection and tracking optimized for embedded hardware.

    YOLOOpenCVNVIDIA JetsonTypeScript
    2025Computer Vision
    Live ↗
  • 05

    Chest X-Ray Pneumonia Detection

    Full-stack medical AI for pneumonia detection from chest X-rays. EfficientNetB0 → ONNX, FastAPI backend, React frontend.

    EfficientNetB0ONNXFastAPIReactTypeScript
    2025Medical AI
    GitHub ↗
  • 06

    Proxora Global LLC

    Founded and led an IT consulting firm delivering AI-powered digital transformation solutions to 15+ enterprise clients across Uzbekistan.

    AI StrategyMLOpsConsulting
    2024–2025Venture

About

I'm a Machine Learning Engineer with 5+ years of experience building production AI trusted at national scale. My work lives at the intersection of computer vision, biometrics, and edge AI — engineering systems that operate reliably under real-world constraints, not just in notebooks.

At UZINFOCOM LLC, I built MyID and PalmID — biometric platforms embedded in Uzbekistan's banking infrastructure. That meant adversarial conditions, regulatory compliance, encrypted model deployment on IoT devices, and latency budgets measured in milliseconds.

I hold degrees in Software Engineering (TUIT) and Artificial Intelligence(Innopolis University), and have won three national and international AI competitions. Today I work at AI Implementation Group and build side projects that push the boundaries of what's possible.

  • CurrentAI Implementation Group
  • LocationTashkent, Uzbekistan
  • EducationTUIT + Innopolis University
  • GitHub@bnutfilloyev
  • Repos74 public
  • Followers91
  • AwardWorld AI & Data Challenge
  • AwardOpen Data Challenge UZ
  • AwardVoice AI Challenge 2021
  • Phone+998 88 300 10 01

Skills

Production-focused
Biometric AI
  • ArcFace / FaceNet
  • Liveness Detection
  • Palm Vein Recognition
  • Anti-Spoofing
  • YOLO
  • OpenCV
Edge & Optimization
  • NVIDIA Jetson
  • CUDA / TensorRT
  • ONNX / ONNX Runtime
  • Model Quantization
  • AES-256 Encryption
  • C/C++ SDK
ML & Frameworks
  • PyTorch
  • TensorFlow / Keras
  • Scikit-learn
  • LangGraph / LangChain
  • RAG / Vector Search
  • Python / C/C++
MLOps & Infra
  • Docker / CI/CD
  • MLflow / PyTorchserve
  • FastAPI
  • Qdrant / FAISS
  • ElasticSearch
  • PostgreSQL

Experience

2021–present

Work History

  • January 2025 — Present

    Machine Learning Engineer

    AI Implementation Group

    Architect and deploy production ML models for enterprise-scale AI solutions. Design MLOps pipelines for automated model training, testing, deployment, and monitoring.

  • November 2021 — January 2025

    Machine Learning Engineer & Researcher

    UZINFOCOM LLC

    Engineered MyID facial recognition (50+ banks, 10M+ monthly requests, 99.7% accuracy). Built PalmID C/C++ SDK with 60% inference speedup via TensorRT. ArcFace/FaceNet on edge with 8× memory reduction.

  • January — September 2025

    Lead AI Engineer

    Robocontest.uz

    Computer vision for autonomous robotics platform supporting 500+ participants. Real-time object detection optimized for embedded hardware.

  • September 2024 — November 2025

    CEO & Founder

    Proxora Global LLC

    Founded IT consulting firm, AI-powered digital transformation for 15+ enterprise clients.

  • May — August 2021

    Computer Vision Engineer

    TASS Vision

    Object detection on NVIDIA Jetson at 30 FPS. 45% latency reduction via CUDA optimization.

Education

  • 2020 — 2024

    B.Sc. Software Engineering

    Tashkent University of Information Technology

  • 2022 — 2023

    B.Eng. Artificial Intelligence

    Innopolis University

Awards

  • World AI & Data Challenge

    Winner

  • Open Data Challenge Uzbekistan

    Winner

  • Voice AI Challenge 2021

    Winner

Full resume available

PDF — detailed experience, projects and references.

Download PDF ↓

Contact

Let's talk

Open to new projects,
collaborations & full-time roles.

Send a message