From Rookie to Tech Pro in One Year: How Two New Engineers Mastered Real-World Projects

Growth Process

New hires are assigned tasks in a tiered manner, progressing from simple, well‑bounded assignments to complex, cross‑team projects. Early tasks are paired with detailed SOP documents, business‑process diagrams, and one‑on‑one mentorship from senior engineers and team leads (TL). The TL helps decompose difficult problems, validates feasibility, and provides iterative feedback at critical milestones.

Project Examples

Model Inference Optimization

The algorithm engineer optimized deployment models used for algorithm training. Initial performance gaps were identified through TL‑guided analysis. External research papers and industry examples were consulted, leading to multiple optimization versions. The final version met runtime requirements, then reduced response time by 50% and doubled throughput.

AI‑Generated Test Case Initiative

The test‑development engineer owned the “AI‑generated test case” project. Responsibilities included defining project goals, expected benefits, detailed implementation plans, operation schedules, and milestones. Meeting agendas were prepared in advance, covering objectives, progress, required support, and timelines. Coordination with the efficiency platform and other departments resolved business‑rule ambiguities. A post‑release retrospective identified issues, causes, and improvement actions, establishing a feedback loop for continuous enhancement.

DeepSeek R1 Model Optimization

When DeepSeek R1, a large open‑source model with novel structures, was introduced, the engineer split the optimization into stages based on difficulty. Internal research was conducted while monitoring industry‑wide optimization techniques. Over three optimization releases, inference latency was halved and throughput doubled, satisfying and then exceeding business real‑time (RT) requirements.

AI Automation Testing

For the full‑process AI‑automated testing project, the engineer selected technologies based on problem context, performed proof‑of‑concept (POC) work, and secured resource support without being forced into a specific stack. The team tolerated early failures, focusing on extracting valuable lessons and iterating. Successful prototypes were shared in internal tech talks and quality platform sessions, providing a venue for demonstration and knowledge exchange.

Support Mechanisms

SOP and Documentation: Detailed SOPs and business‑process diagrams enable rapid comprehension of complex systems.

Mentorship: Senior engineers provide real‑time guidance, reducing time spent on undocumented procedures.

Project Management Tools: Standardized processes and tools ensure transparent information flow, task tracking, and document preservation.

Feedback and Retrospective: Structured post‑mortems focus on problem identification, root‑cause analysis, and actionable improvements.