Haozheng Li 李浩正

Computer science undergraduate working at the intersection of AI prototyping, technical solutions, and enterprise customer needs.

About Me

Hometown:: Nanning
Location:: Shenzhen / Beijing, China
Target Roles:: Solutions Intern, Pre-sales Technical Support Intern, or B2B Sales Intern / Sales Assistant
Industry Focus:: AI, enterprise software, and digital transformation
Human Languages:: English, Mandarin, Cantonese
Interests:: Embodied AI, artificial intelligence, software
Email:: [email protected]

B.Eng. Candidate in Computer Science and TechnologySep 2023 – Expected Jun 2027

School of Information Engineering, Minzu University of ChinaBeijing, China

AI Solutions Intern (B2B / Pre-sales Support)Feb 2026 – May 2026

Foxu AI (Shanghai) Technology Co., Ltd. / China Stone Management Consulting GroupBeijing, China

Supported discovery, solution design, and prototype demonstrations for pharmaceutical and chemical clients, translating business workflows and pain points into AI CRM and report-automation capabilities.
Built an end-to-end GLP report automation POC with LangGraph and FastAPI, using Examiner, Corrector, and Reviewer agents for document generation, data validation, and human review.
Structured customer leads, profiles, business forms, and data fields, and helped align client needs with product and engineering deliverables.

B2B Vendor Liaison and Procurement CoordinatorDec 2023 – Apr 2024

Guangxi Zhanqi Co., Ltd.Guangxi, China

Scoped office-equipment requirements, compared vendor proposals, coordinated specifications, quotations, delivery, and deployment, and helped reduce procurement costs by more than 10%.

Physics-Guided Dual-Magnet Localization for Robotic SensingJun 2025 – May 2026

Undergraduate Research Contributor, The Chinese University of Hong KongHong Kong SAR, China

Contributed to a physics-guided real-time dual-magnet localization study for robotic sensing and magnet-assisted medical interventions.
Conducted sensor-array experiments on a magnetic sensing platform, evaluated real-world data quality, analyzed challenging cases, and summarized experimental observations to support model refinement.
Evaluated data-efficient simulation-to-reality adaptation by comparing synthetic pretraining and real-data fine-tuning settings, analyzing physics-based field reconstruction and residual patterns.
Achieved 5.30 ms inference latency and mean position errors of 2.45/2.53 mm under the 60% fine-tuning setting; co-first-authored manuscript under revision at IEEE Transactions on Instrumentation and Measurement.

Multimodal Motion Assessment and Feedback SystemOct 2023 – Mar 2025

Undergraduate Research Assistant, Minzu University of ChinaBeijing, China

Developed and tested a computer-vision workflow for sports posture analysis using YOLO-based detection, human keypoint estimation, and rule-based movement assessment.
Combined model outputs with multimodal AI-agent support to generate automated analysis and interpretable feedback for tennis and related motion scenarios.

FPGA HDL PortfolioVerilog FPGA and MIPS CPU design portfolio

SQL Copilot AgentAI-driven SQL autocompletion and query assistant

GoZero AgentGo AI agent based on supervised learning, reinforcement learning, and MCTS

Oracle Bone Image ProcessingOracle bone image rectification, contour fitting, and character segmentation pipeline