From Campus to Backend Engineer: Growth, System Refactoring, and AI Assistant at JD

In 2021 the author graduated from the University of Chinese Academy of Sciences and joined JD as a new graduate, working on merchant content and AI assistants. The article aims to help fresh graduates adapt to the workplace.

Start from Zero, Steady Progress

Newcomers often feel undervalued, face legacy code, and encounter communication barriers. The author advises building trust, taking small tasks, documenting work, and gradually taking on more responsibilities.

When assigned few development tasks, proactively investigate alerts, logs, and even front‑end work to demonstrate value and accelerate learning.

If existing code seems flawed, consider incremental micro‑refactoring and write documentation to aid future maintenance.

For communication challenges with upstream/downstream teams, suppress emotions, practice empathy, and ask clear questions; eye contact can reduce communication cost.

JD Merchant Conference: First Closed‑Loop Experience

In April 2022 the author’s team faced a high‑traffic live‑streaming demand for a merchant conference. The seven‑year‑old system had an outdated architecture, prompting a domain‑driven redesign.

Key steps included defining business boundaries, controlling change, and using domain modeling to achieve high cohesion and low coupling. The refactor introduced clear domain boundaries, isolated vertical business logic, and shared horizontal capabilities.

Define boundaries with real business concepts.

Control change by extracting variation points and isolating vertical services.

Examples: unify terminology, abstract business growth, and partition problems into domains.

Before the refactor the architecture suffered from tangled code, unclear responsibilities, and fragmented deployments. After refactoring, the system achieved clear domain boundaries and controlled variation, improving maintainability and scalability.

Reject Restlessness, Pursue Perfection

The author reflects on the stereotype of programmers as short‑term workers, urging humility, continuous questioning, and deep technical exploration beyond simple CRUD tasks.

Hard problems, such as reading Tomcat source code, build resilience and reveal the beauty of technology.

Large‑Scale Promotion Preparation (618/Double‑11)

Each year JD prepares months in advance for major sales events, iterating on core processes, identifying weak points, optimizing, and conducting stress tests. Continuous improvement and post‑mortem analysis are essential to handle unexpected failures.

Continuous Learning and Self‑Innovation

Backend engineers should define personal technical roadmaps, recognize that formal education ends at graduation, and pursue ongoing learning through practice, reading source code, and mastering protocols like HTTP in real scenarios.

JD Merchant AI Assistant: Breaking Boundaries

Following the rise of ChatGPT in 2022, JD identified merchant pain points such as fragmented manuals, low‑quality intelligent客服, and complex cross‑system operations. A dedicated large‑model team built a general‑purpose assistant platform, iterating over three quarters to reach a 3.0 version with gray‑release, permission control, rate limiting, and full‑stack monitoring.

Summary

The author emphasizes steady growth, domain‑driven system design, meticulous promotion preparation, and relentless learning as keys to becoming a competent backend engineer.

Recruitment Notice

JD is actively hiring for campus and social recruitment positions, including PC client engineers and app developers across Android, iOS, PC, and web, as well as AI assistant roles. Interested candidates can email [email protected].