Umidjon Abdukhamidov
Head of the AI solutions group for internal banking services. Over 5 years of experience across IT and banking: from backend development to business analysis and AI system deployment. Turns business requirements into working solutions - from specifications to production launch.
About
Umidjon Abdukhamidov is a specialist with over 5 years of experience at the intersection of software development, business analysis, and applied AI.
He began his career as a Python developer, building website backends with Django, Telegram bots, and corporate systems. He then moved into business analysis at Fido-Biznes, where he handled data collection and analysis, task management for developers, and document workflow support.
In 2024, he joined Uzpromstroybank (SQB): analyzing requirements for the Joyda mobile app, designing integrations and API specifications, launching new banking products, and automating processes in the CROBS core banking system.
After moving to the AI Department, he took on leadership of the Bank Internal Services Group. He manages the full lifecycle of AI systems for the bank's internal operations: gathering requirements from departments, drafting specifications, preparing and labeling data, fine-tuning NLP models (BERT, LLM), and integrating solutions into banking infrastructure.
His core strength is working equally well with department leadership and technical teams - every project travels the full path from problem identification to production launch.
Achievements
- Contributed to the SQB Mobile (ex. Joyda) app: API specifications, integrations, launch of new products
- Automated ATM transaction accounting and cash flow monitoring in CROBS
- «Obnal» project - deployed an AI model for automatic classification of ambiguous payment purposes
- Designed and deployed «Ijro AI» - an automatic routing system for incoming bank correspondence using OCR + BERT + LLM, integrated with an external vendor and internal bank systems
- Built a PDF document comparison tool for the Operations Department using Surya OCR + PaddleOCR
- Built AI-powered Telegram assistants for SQB's Audit and Operations Departments
- Contributed to the SQB Mahalla platform - district-level analytics across Uzbekistan with an integrated AI financial advisor
- Developed an AI resume screening system for internal HR: PDF/DOCX analysis, 0-100 scoring, interviewer routing
Skills
Projects
SQB Mahalla
A single digital platform: regional socio-economic analytics plus a mobile workplace for mahalla bankers, with a built-in AI Advisor that turns a business idea into a plan and matching SQB financing.
SQB AI Ijro
AI-powered intake and routing for incoming bank correspondence - OCR, LLM and embedding-based dispatch to responsible departments.
ULAB
A benchmark platform for Uzbek-language AI models: an objective ranking of 17 models across 1,377 questions via an interactive dashboard.