Zhangir Meirbek -_-

Designed and documented a Python benchmarking pipeline for evaluating theory of mind in LLMs, improving evaluation coverage by 40% across 1,000+ test cases. Engineered a SQL-backed versioning system to reduce labeling errors and communicated findings via technical write-ups.

Led system architecture and engineering coordination for a 6-person national robotics team, integrating mechanical, electrical, and software workstreams. Designed and built the robot chassis and intake system, achieving 17th place among 191 countries at the FIRST Global Challenge.

Owned the complete SDLC for an EdTech platform, from PostgreSQL schema design through Next.js 15 front-end to cloud deployment with automated build pipelines. Built a normalized PostgreSQL-backed knowledge graph with an SVG front-end, deployed via CI/CD and merged through peer-reviewed pull requests.

Built a functional ESP32-based portable music player prototype by integrating SPI/UART peripherals, a TFT display, rotary encoder controls, and microSD-based audio playback. Designed a responsive embedded UI with playback controls and themes, and debugged hardware/firmware integration issues.

Collaborated with a 4-person agile team to prototype a RAG pipeline that enriched sparse user profiles with relational data, increasing recommendation relevance by 23% (simulated CTR). Designed a hybrid retrieval module to reduce cold-start failures by 30% and optimized the NLP inference layer for sub-200ms REST API responses.

Pursuing a Bachelor of Applied Science in Electrical Engineering with a minor in Computer Science, expected to graduate in May 2029.

Completed the Supervised Machine Learning: Regression and Classification course offered by DeepLearning.AI and Stanford Online.

Completed the Solutions Architecture Job Simulation offered by AWS through Forage.