Wang Jianjun

Internships

My work on backend data services, design agents and enterprise AI: the problems I worked on and how I contributed.

Tencent 腾讯

Tencent

Current internship

Backend Developer · Internship · Shenzhen, China

Oct 2026 – Present

Monitoring data services

Contribute to backend services for cross-platform monitoring, covering ingestion, analytical queries and usage metering, with a focus on data consistency and reliability.

  • Ingestion and error handling

    Contribute to collection services and integration workflows. Clarify input validation, error feedback and overload handling, and support integration testing and debugging to make ingestion easier to verify and maintain.

  • Queries and statistical consistency

    Develop query interfaces and reporting logic, checking that category breakdowns, totals and presentation use consistent definitions. Investigate discrepancies and query-efficiency issues through validation and edge-case review.

  • Usage metering and reconciliation

    Contribute to usage statistics and data checks. Review day boundaries, repeated execution and state changes, and add regression checks with attention to accuracy and traceability.

Data ingestionStatistical consistencyService reliability
快手 Kuaishou

Kuaishou

Agent Developer · Internship · Beijing, China

Aug – Oct 2026

Design agents and delivery evaluation

Built evaluation workflows and contributed to editing capabilities and tool integration for agents that turn requirements into editable designs, with a focus on verifiable outputs and iterative improvement.

  • Design-agent evaluation

    Built evaluation workflows that translated requirements into checkable acceptance criteria. Inspected design outputs and analyzed failure cases, preserving run evidence for comparable regression checks after fixes.

  • Design-content editing

    Contributed to agents’ ability to read and edit design content. Focused on whether generated content remained editable and changes matched the request, and helped test complex edits and error cases.

  • Tool integration and debugging

    Contributed to tool-based access to generation capabilities and integration testing. Investigated timeouts and connection failures, distinguished execution status from delivered results, and improved error feedback and user workflows.

Agent evaluationDesign toolsDiagnosis & regression
蕉内 Bananain

Bananain

AI Application Engineer · Internship · China

Mar – Jul 2026

Visual production and enterprise AI tools

Worked on product workflows, tool development and application support for visual content production and internal AI applications, bringing model capabilities into everyday work.

  • Product visuals and design plugins

    Developed visual-generation tools and design plugins, translating content needs into generation, editing and review workflows. Focused on editable results and how the tools fitted designers’ daily work.

  • Enterprise AI integrations

    Contributed to model, tool and business-capability integrations. Defined application workflows and data-access permissions to support reuse across scenarios while clarifying tool-use and access boundaries.

  • Requirements and user support

    Contributed to internal AI requirements and platform selection. Documented capabilities, limitations and usage instructions, and supported feature validation and training to help teams understand and use the tools.

Visual generationDesign workflowsAI integrations

Shenzhen Linyuan Technology

AI Application Engineer · Internship · China

Oct 2025 – Feb 2026

E-commerce visual-generation applications

Contributed to e-commerce visual-generation applications, connecting style analysis, model invocation and generation tasks across backend services, task interaction and output validation.

  • Visual style and generation requirements

    Analyzed visual references and generation requirements, organizing style characteristics into reusable prompts and references to support generation and revision for different content needs.

  • Model integration and task interaction

    Contributed to model integrations and generation-task services, connecting submission, progress updates and result retrieval with attention to waiting states and error handling.

  • Output evaluation

    Reviewed generated outputs and categorized issues. Combined model-based evaluation with human review to guide prompt and workflow revisions, focusing on style alignment and usable results.

Model integrationGeneration tasksOutput evaluation

Public projects and university collaborations are listed separately.

Explore public projects