Full-Stack AI Engineer · Dots

Job Responsibilities

1. End-to-End Function Development & Delivery: Participate in full-stack R&D for the Dots AI conversational application. Leverage AI programming tools (e.g., Claude Code, Cursor, Codex) for human-AI collaborative programming, and independently deliver end-to-end architecture design (Client-Server-AI) and high-quality implementation for complete functional modules.
2. Agent System & Workflow Development: Contribute to building Agent systems that connect large language models (LLMs) with business scenarios. Design and implement core mechanisms including multi-step reasoning, dynamic workflow orchestration, multi-model routing, diverse tool calling, and DeepResearch.
3. Server-Side & High-Concurrency Conversation System Development: Involve in core server-side architecture design for AI conversational products. Build streaming message distribution and routing systems supporting tens of millions of concurrent connections based on SSE, WebSocket, gRPC and other protocols. Drive capacity planning, full-link tracing, and performance bottleneck optimization for microservice systems.
4. Front-End & Cross-Platform Experience Optimization: Own cross-platform technical vision, deeply engage in core development for major front-end platforms (iOS / Android / Web / RN). Tackle challenges in rich text and complex card rendering, multi-modal interaction (audio, video, images), sophisticated animations, and extreme client performance optimization (startup, memory, lag, and smoothness).
5. Full-Link High-Availability Assurance: Design and implement robust engineering fallback mechanisms to address the uncertainty and instability of complex AI tasks. Build highly available service architectures covering intermediate state storage and recovery, long-link fault tolerance, intelligent retry, and service degradation to ensure industrial-grade stability for core scenarios.

Job Requirements

• Full-Stack Foundation & Motivation: Strong motivation or practical experience in full-stack development. Proficient in at least one back-end language (Java / Go, etc.), or familiar with one modern front-end framework (React / Vue), or possess extensive experience in mobile (iOS / Android) architecture tuning.
• Agent Architecture Expertise: In-depth understanding of LLM working principles and capability boundaries. Hands-on experience in building complex Agent systems or long-horizon task flows. Familiar with task decomposition, context enhancement, and intermediate state management.
• High-Availability & High-Concurrency Architecture: Development experience in high-concurrency long-connection systems and microservice architectures. Clear solutions and practical experience in distributed system stability governance, full-link tracing, disaster recovery, and service degradation; capable of independently resolving complex system engineering bottlenecks.
• Senior AI Product User: Strong interest in AI products, with intensive and frequent usage of mainstream LLMs (ChatGPT, Claude, Gemini, Kimi) and AI programming tools (Cursor, Copilot, Windsurf). Able to objectively evaluate the engineering architecture strengths, weaknesses, and applicable scenarios of these products.
• Self-Driven & Ownership Mindset: Proactively decompose ambiguous AI business problems and drive cross-role closed-loop delivery. Excellent systematic thinking, technical abstraction, and strong learning agility.

Preferred Qualifications

• Experience in end-to-end architecture delivery of AI products from 0 to 1, or independent design and launch of complete AI tools / products; capable of independently delivering full functional modules end-to-end using AI tools.
• R&D experience in industry-leading AI conversational or native products with over 1 million DAU, with proven expertise in architecture evolution, high-concurrency governance, and performance tuning for massive real-user and complex multi-modal scenarios.
• Active contributor to well-known open-source projects, especially in AI infrastructure, LLM application frameworks, or Agent orchestration frameworks; in-depth insights into LLM capability boundaries.