Robotics / software engineer with 10+ years of experience and an open-source contributor to Nvidia IsaacLab, Go2 Omniverse SDK, and the Unitree Go2 SDK. I design end-to-end autonomous systems — perception, locomotion, and control — and build resilient backend & fintech infrastructure that runs in production.
Experience
10+ years across robotics R&D, autonomy, machine learning, and fintech engineering.
Next-gen payments & tokenized loyalty platform
Led engineering for a customizable payments/coupon platform. Launched an L2 coin on Etherlink (Tezos) for a loyalty program serving Switzerland's largest health insurer — reaching 42,355 holders and 167,819 on-chain transfers, ranking #2 by holders in the Tezos (Etherlink) network. Source ↗
Grew from Senior Engineer → Dev Lead → CTO
Established two American neobanks from the ground up and built production-grade infrastructure for complex financial environments, leading engineering teams across the US and Europe over a multi-year operation.
Memory-driven visual navigation & robotic swarms
Co-authored research on MLLM memory-driven visual navigation (Unitree GO1) and blockchain-managed robotic swarms with Hybrid ACO. Work featured by Texas A&M Engineering ("Smarter than your average dog").
Frontier AI & Robotics
Consulted on advanced robotics problems alongside leading experts in the field.
Part-time job
Taught 1000+ students in robotics, machine learning, algorithms & data structures, mobile development, system reliability, and operating systems.
End-to-end robotic welding on the edge of science
RPA & automation at Kazakhstan's largest bank
Built a commercially successful RPA studio (8× faster robot development) and a Forex trading robot approving 100+ transactions/day (~$100M), cutting approval time from 15 to 2 minutes.
Computer vision & industrial automation
Built a computer-vision system for bitumen quantity control (−30% material use), a semi-automatic filling conveyor (3.6× throughput), and industrial steam (1.2 MW) & thermal-oil (420 kW) boilers from scratch.
Machine learning & humanoid balancing research
Created and taught "Introduction to Machine Learning"; supervised lab AI projects including a humanoid balancing on an unstable platform using neural networks.
Engineering work
Selected systems I've designed and built — from academic research to commercial robotics and industrial engineering.
Converted a $1,600 Unitree Go2 Air into a wheeled version — a configuration that normally costs around $16,000 from the manufacturer. Built and validated in Nvidia Isaac Lab, turning a legged platform into a high-mobility wheeled robot for faster traversal on flat and urban terrain.
Master's thesis work: designed and built a biped humanoid robot from scratch and developed a balancing control algorithm powered by a neural network — without PID controllers. The system learns to stabilize the platform on an unstable surface, combining mechanical design, embedded control, and machine-learning-based balance recovery in real time.
Bachelor's diploma project: designed and built a fully automated mini-warehouse with autonomous mobile robots — inspired by Amazon warehouse robotics. The system coordinates multiple robots for item storage, retrieval, and transport across the warehouse floor, demonstrating end-to-end automation from task scheduling to physical execution.
Designed and assembled industrial steam boiler systems from scratch — including a 1.2 MW steam boiler and a 420 kW thermal-oil boiler. Full cycle: thermal engineering, mechanical layout, piping, electrical control panels, and on-site commissioning. Saved around 60% of the budget compared to purchasing ready-made units from manufacturers.
A last-mile delivery robot built from scratch — hardware platform plus a Level-4 autopilot written and validated in the CARLA simulator using Unreal Engine 4. Engineered for the lowest bill-of-materials cost on the market while keeping an all-weather perception stack.
A legged delivery platform that moves seamlessly between indoor and outdoor environments, climbs steps, and handles constrained spaces that stop wheeled robots. Built on a Unitree Go2 chassis with a modular sensor array.
Reinforcement-learning and control experiments for real-world interaction tasks such as autonomous door opening — perception, manipulation, and locomotion trained in Nvidia Isaac Lab.
An end-to-end robotic beam-welding system that boosts production speed by 1.5× and cuts defects by 60% compared to human welders. Included an award-winning 3D analyzer for Tekla (Trimble), integrated with ROS.
🔒 There are more interesting projects under NDA that I can't show publicly — only the work I'm allowed to share is listed here.
Education
Hangzhou Dianzi University, China
Al-Farabi Kazakh National University, Almaty
Research
Peer-reviewed and preprint work in robotics, autonomy, and multi-robot systems.
ETASR — Engineering, Technology & Applied Science Research
A survey of SLAM for off-road mobile robots, categorizing LiDAR-based, visual, multi-sensor-fusion, and learning/semantic approaches. Analyzes accuracy, drift, robustness, and efficiency, and identifies research gaps toward hybrid, computation-aware SLAM and standardized off-road benchmarks.
View publication ↗Texas A&M University · ResearchGate
A memory-based navigation framework for terrestrial robots using a customized multimodal LLM to interpret visual inputs and generate navigation commands. A Unitree GO1 reuses previously traversed paths to reduce re-exploration, with a hybrid deliberation + reactive controller improving path retracing.
View publication ↗ResearchGate
A dynamic multi-robot mapping framework combining blockchain-based swarm contracts for decentralized task allocation with a Hybrid Ant Colony Optimization algorithm for path planning. Deployed LiDAR-equipped Unitree Go2 robots in cooperative and competitive modes to map sub-areas with verifiable full coverage.
View publication ↗Contact
Open to robotics engineering, autonomy, and applied-ML roles and collaborations. The fastest way to reach me is email.