Computer Science graduate · AI and software engineer
I am Chenyu Zhang, a Computer Science graduate building AI applications and production-oriented software systems. I am currently seeking full-time research assistant and AI engineering opportunities where strong implementation, reliable evaluation, and real-world delivery matter.
Selected work
A public, evidence-first research workbench that connects document retrieval, citation verification, MCP tool calling, version comparison, human-review routing, and auditable events in one reproducible workflow.
A bounded autonomous underwriting and capital-action system for invoice-backed financing. Eight specialized agents turn public evidence into a risk credential, enforce deterministic vault policy, cap executable principal, and anchor both the credential and execution intent on Casper Testnet.
An evaluation system for auditing the safety behavior of embodied LLM agents. The current artifact set includes a trace schema, safety taxonomy, rule baseline, evaluators, report generation, an AI2-THOR adapter, and a technical report draft.
A local-first confidential contract intelligence workspace for consumer Windows hardware. It combines QVAC local inference with deterministic policy checks, citation-bound findings, amendment drafting, prompt-injection defenses, and auditable report export without remote AI APIs.
Technical direction
Turning model capabilities into usable systems through retrieval, tool calling, databases, deployment, and clear human-review boundaries.
Building observable and reproducible agent workflows that interact with tools, documents, services, and external state.
Evaluation methods that make evidence provenance, failure modes, model judgments, and operational boundaries explicit.
Applying software and AI engineering to scheduling, simulation, digital twins, robotics, and other systems linked to physical operations.
Verification over assertion
The portfolio distinguishes finished artifacts from active research. Public code, deployed systems, CI results, Testnet records, and documented evaluation protocols are linked where available. EvidenceBridge is publicly testable without an API key; SafeTrace experiments and the first-author manuscript remain explicitly in progress.
Open GitHub profile