主頁Home 產品服務Services 關於我們About 招聘Careers 聯絡我們Contact
🚀 AI 驅動 · 未來已至 🚀 AI-Powered · Future is Now

智能科技
重塑商業未來 Intelligent Tech
Reshaping Tomorrow

ASAIR 專注於 AI 系統的研發與落地，為企業提供從設計到部署的一站式智能化解決方案，助力業務持續增長。 ASAIR specializes in AI system development and deployment, delivering end-to-end intelligent solutions that drive sustainable business growth.

探索服務 Explore Services → 立即聯絡 Get in Touch
AI Illustration
50+
企業客戶Enterprise Clients
200+
已交付項目Projects Delivered
99.9%
系統可用性System Uptime
24/7
技術支援Technical Support
我們的服務Our Services

全方位 AI 技術服務 Full-Stack AI Solutions

無論是定制開發、系統維護，還是互聯網軟件與專業諮詢，我們都能為您的業務提供精準匹配的 AI 技術支持。 From custom AI development and system maintenance to internet software and expert consulting, we deliver precision-matched AI support for your business.

🧠

AI 系統開發AI System Development

從需求分析到模型訓練，再到系統集成與上線，提供全週期 AI 開發服務，涵蓋機器學習、自然語言處理與計算機視覺等核心技術領域。 From requirements analysis to model training and system integration, we provide full-lifecycle AI development covering machine learning, NLP, and computer vision.

Machine Learning NLP Computer Vision
⚙️

AI 系統維護與銷售AI Maintenance & Sales

持續監控、性能調優與版本迭代，確保 AI 系統穩定運行。同時提供成熟的 AI 產品及解決方案，滿足不同規模企業的採購需求。 Continuous monitoring, performance tuning, and version iteration keep your AI systems running reliably. We also offer proven AI products and solutions for enterprises of all sizes.

DevOps MLOps 24/7 Support
🌐

互聯網軟件服務Internet Software Services

提供高可用的 Web 應用、SaaS 平台及 API 服務開發，助力企業快速構建數字化業務生態，打通各端數據鏈路。 We build high-availability web applications, SaaS platforms, and API services to help enterprises rapidly construct digital business ecosystems and connect data pipelines.

SaaS API Integration Cloud
💡

軟件諮詢服務Software Consulting

深度解析業務痛點，提供 AI 技術選型、系統架構設計及數字化轉型路線圖，幫助企業做出最具價值的技術決策。 We analyze business challenges in depth and provide AI technology selection, system architecture design, and digital transformation roadmaps to support high-value technology decisions.

Strategy Architecture Roadmap
✦ 我們的使命Our Mission

以 AI 之力
驅動無限可能Powering Infinite
Possibilities with AI

我們相信技術應服務於人，致力將複雜的 AI 能力轉化為簡單易用的商業工具，讓每一家企業都能受益於智能化的力量。 We believe technology should serve people — transforming complex AI capabilities into accessible business tools so every enterprise can harness the power of intelligence.

98%
客戶滿意度Client Satisfaction
5+
年行業經驗Years Experience
120+
🤖 AI 模型部署🤖 AI Models Deployed
關於 ASAIRAbout ASAIR

技術領先，
人為本心 Technology First,
People at Heart

ASAIR 由一支融合 AI 研究員、工程師與商業顧問的多元團隊組成。我們深耕 AI 行業多年，以務實的技術態度和以客戶為中心的服務理念，持續為合作夥伴創造可量化的業務價值。 ASAIR is built by a diverse team of AI researchers, engineers, and business consultants. With years of deep industry expertise and a client-centric approach, we consistently deliver measurable business value to our partners.

頂尖研發團隊Elite R&D Team

匯聚機器學習、深度學習及軟件工程領域的資深專家，持續推動技術創新與突破。Senior experts in machine learning, deep learning, and software engineering, continuously driving technological innovation.

定制化 AI 解決方案Custom AI Solutions

根據企業實際業務場景量身定制 AI 方案，確保技術落地真正創造商業價值。Tailored AI solutions designed around your actual business scenarios, ensuring technology deployment creates real commercial value.

全程技術支援End-to-End Support

從項目啟動到系統上線及後續運維，全程提供專業技術團隊陪伴與保障。From project kickoff through system launch and ongoing maintenance, our professional team supports you every step of the way.

合作流程How We Work

四步驟快速落地 Four Steps to Go Live Fast

01

需求評估Assessment

深入了解業務目標與技術現狀，識別 AI 落地的最佳切入點。We dive deep into your business goals and technical landscape to identify the best entry points for AI implementation.

02

方案設計Design

基於評估結果制定技術架構與實施計劃，確保方案可行且具有擴展性。Based on the assessment, we craft a technical architecture and implementation plan that is both feasible and scalable.

03

開發部署Development

採用敏捷開發模式快速迭代，按期交付高質量 AI 系統並完成生產部署。Using agile development for rapid iteration, we deliver high-quality AI systems on schedule and deploy them into production.

04

維護優化Optimization

上線後持續監控性能指標，定期優化模型與業務邏輯，保障長期穩定運行。After go-live, we continuously monitor performance metrics and optimize models and business logic to ensure long-term stable operation.

加入我們Join Our Team

誠邀精英加入 We're Hiring

我們正在尋找對 AI 充滿熱情的工程師與技術人才，共同構建下一代智能 Agent 平台。如果你熱愛技術、渴望挑戰，歡迎加入 ASAIR。 We're looking for passionate engineers and technologists to help build the next generation of intelligent Agent platforms. If you love technology and thrive on challenges, we'd love to hear from you.

🛠️

AI Agent 工程師（全棧方向） AI Agent Engineer — Full-Stack

中級Mid-level 約 3 年經驗~3 years exp.
▾
岗位职责
  • 参与 AI Agent 平台与智能客服产品的端到端研发，覆盖前端交互、后端服务及 Agent 编排全链路。
  • 基于 LLM Gateway 与知识库引擎，开发 Agent 应用层能力：多轮会话、工具调用（Function Calling / MCP）、任务编排与人工接管。
  • 负责 Agent 运行时的可观测性建设：链路追踪、Token 成本核算、效果回流与评测面板。
  • 对接内外部系统（工单、CRM、IM 渠道等），将 Agent 能力完整落地为可交付的产品功能。
  • 参与多租户 SaaS 化改造，保障系统并发性能、稳定性与租户数据隔离。
任职要求
  • 约 3 年后端或全栈研发经验，计算机相关专业本科及以上学历。
  • 熟练掌握 Java（Spring Boot 3.x）或 Python（FastAPI），至少一门为主力语言，另一门可读可改。
  • 具备扎实的前端开发能力，熟悉 React / Vue + TypeScript，可独立交付中后台界面与对话式 UI。
  • 有 LLM 应用开发实战经验：OpenAI 兼容 API、流式响应（SSE）、Prompt 工程、Function Calling / MCP。
  • 熟悉 PostgreSQL、Redis，了解向量检索（pgvector / Milvus 等）基本原理与使用。
  • 熟悉 Docker，了解云上部署（AWS ECS / Kubernetes）及 CI/CD 流程。
加分项
LangGraph / Dify / Coze 多租户 SaaS 产品经验 金融 / Web3 / 客服 SaaS 背景 开源贡献 / 个人 Agent 项目
Responsibilities
  • Own end-to-end development of our AI Agent platform and intelligent customer-service product, spanning front-end interaction, backend services, and agent orchestration.
  • Build application-layer agent capabilities on top of our LLM Gateway and knowledge base engine: multi-turn dialogue, tool calling (Function Calling / MCP), task orchestration, and human handoff.
  • Build observability for the agent runtime: distributed trace logging, token cost accounting, feedback loops, and evaluation dashboards.
  • Integrate with internal and external systems (ticketing, CRM, IM channels) to ship agent capabilities as production-ready product features.
  • Contribute to multi-tenant SaaS architecture to ensure throughput, stability, and tenant data isolation.
Requirements
  • ~3 years of backend or full-stack engineering experience; Bachelor's degree in Computer Science or a related field.
  • Strong command of Java (Spring Boot 3.x) or Python (FastAPI) as a primary language, with working proficiency in the other.
  • Solid front-end skills — React / Vue with TypeScript — able to independently deliver admin dashboards and conversational UI.
  • Hands-on LLM application development: OpenAI-compatible APIs, streaming responses (SSE), prompt engineering, Function Calling / MCP.
  • Proficient with PostgreSQL and Redis; familiar with vector search (pgvector, Milvus, or similar).
  • Comfortable with Docker; familiar with cloud deployment (AWS ECS / Kubernetes) and CI/CD pipelines.
Nice to Have
LangGraph / Dify / Coze Multi-tenant SaaS experience Finance / Web3 / CX SaaS background Open-source contributions
📧 請將簡歷發送至：[email protected] 📧 Send your resume to: [email protected] 立即申請 → Apply Now →
🧠

AI Agent 工程師（模型 / 算法方向） AI Agent Engineer — Model & Algorithm

中高級Mid-to-Senior 3–5 年經驗3–5 years exp.
▾
岗位职责
  • 负责 Agent 核心"大脑"的算法设计与持续迭代：意图识别、检索增强生成（RAG）、多跳推理、规划与反思机制。
  • 优化检索全链路：文档切分策略、Embedding 选型与版本管理、混合检索（向量 + BM25）、Rerank 精排。
  • 建设 Agent 评测体系：构建评测数据集，设计自动化评估指标（准确率、幻觉率、工具调用成功率），驱动效果闭环迭代。
  • 设计多语言场景下的意图识别与检索策略，保障跨语种效果一致性。
  • 负责模型选型与成本效益平衡：多模型路由策略、蒸馏与微调（SFT / LoRA）、推理加速优化。
  • 与工程团队深度协作，将算法方案落地为可上线、可监控的生产级服务。
任职要求
  • 3–5 年算法或 AI 应用研发经验，计算机 / 数学 / 统计等相关专业硕士优先。
  • 精通 Python，熟悉 PyTorch 及 HuggingFace 生态系统。
  • 有 RAG 系统的完整落地经验，深刻理解切分、召回、重排、生成各环节的失效模式与调优方法。
  • 熟悉主流 LLM 的能力边界，有 Prompt 优化与结构化输出的实战经验。
  • 具备严谨的评测思维，能够以数据驱动而非主观直觉推进效果迭代。
  • 熟悉向量数据库原理与检索基础（ANN 搜索、相似度度量等）。
加分项
模型微调 (SFT / DPO / LoRA) Agent 规划 / 多 Agent 协作 NLP 顶会论文 / Kaggle 竞赛 金融 / 风控 / 客服领域经验
Responsibilities
  • Design and iterate on the agent's core reasoning layer: intent recognition, retrieval-augmented generation (RAG), multi-hop reasoning, planning, and reflection mechanisms.
  • Optimize the end-to-end retrieval pipeline: document chunking strategy, embedding selection and version management, hybrid retrieval (vector + BM25), and reranking.
  • Build the agent evaluation framework: curate evaluation datasets, define automated metrics (accuracy, hallucination rate, tool-call success rate), and drive a closed improvement loop.
  • Design intent recognition and retrieval strategies for multilingual scenarios to ensure cross-language consistency.
  • Balance model capability against cost: routing strategies across models, distillation and fine-tuning (SFT / LoRA), and inference optimization.
  • Partner closely with engineering to ship algorithmic work as production-grade, monitorable services.
Requirements
  • 3–5 years in algorithm or applied AI engineering; Master's in Computer Science, Mathematics, Statistics, or a related field preferred.
  • Expert-level Python; strong familiarity with PyTorch and the HuggingFace ecosystem.
  • Proven end-to-end RAG system delivery, with deep understanding of failure modes and tuning levers across chunking, retrieval, reranking, and generation.
  • Practical grasp of the capability boundaries of mainstream LLMs; experienced in prompt optimization and structured output.
  • Rigorous evaluation mindset — able to drive iteration with data rather than intuition.
  • Solid understanding of vector databases and retrieval fundamentals (ANN search, similarity metrics, etc.).
Nice to Have
Fine-tuning (SFT / DPO / LoRA) Agent planning / multi-agent systems Top-venue NLP publications / Kaggle Finance / risk / CX domain experience
📧 請將簡歷發送至：[email protected] 📧 Send your resume to: [email protected] 立即申請 → Apply Now →
🔬

AI Agent 測試工程師（質量保障方向） AI Agent Test Engineer — Quality Assurance

中高級Mid-to-Senior 3–5 年經驗3–5 years exp.
▾
岗位职责
  • 负责 AI Agent 产品的质量保障体系建设，覆盖功能正确性、输出效果、性能稳定性及安全合规四个维度。
  • 从 0 到 1 构建 Agent 自动化测试框架：多轮会话回放、工具调用断言、非确定性输出的语义级校验（语义相似度 / LLM-as-Judge）。
  • 与算法团队共建评测集与回归基线，确保每次 Prompt、模型或检索策略变更均可量化、可对比、可回滚。
  • 设计并执行对抗性测试：Prompt 注入、越狱尝试、敏感信息泄露、幻觉诱导等高风险场景验证。
  • 负责流式接口（SSE）与高并发会话场景下的性能压测、稳定性验证与成本评估。
  • 跟踪线上 Bad Case，驱动归因分析与质量闭环，推动问题系统性解决。
任职要求
  • 3–5 年测试开发经验，有自动化测试框架从 0 到 1 构建的完整经历。
  • 精通 Python，能独立设计测试框架与工具链，而非仅使用现成工具。
  • 熟悉接口测试、性能测试工具（Locust / JMeter）及 CI/CD 集成（GitLab CI / Jenkins）。
  • 理解 LLM / RAG 系统的基本原理与典型失效模式（幻觉、召回缺失、工具调用错误等）。
  • 对非确定性系统的质量评测有独立方法论，能接受"没有唯一正确答案"的质量判定场景。
  • 数据意识强，善用指标而非主观感受描述质量状态，能独立输出质量分析报告。
加分项
Ragas / DeepEval / Promptfoo AI 安全 / 红队测试经验 SaaS / 金融行业质量保障经验 熟悉 Java 技术栈
Responsibilities
  • Own the quality assurance strategy for our AI Agent products across four dimensions: functional correctness, output quality, performance stability, and safety compliance.
  • Build an agent test automation framework from the ground up: multi-turn conversation replay, tool-call assertions, and semantic-level validation of non-deterministic output (semantic similarity, LLM-as-Judge).
  • Partner with the algorithm team to build evaluation sets and regression baselines, ensuring every prompt, model, or retrieval change is measurable, comparable, and reversible.
  • Design and execute adversarial testing: prompt injection, jailbreaking, sensitive data leakage, and hallucination-inducing scenario validation.
  • Lead performance, stability, and cost testing for streaming endpoints (SSE) and high-concurrency conversation workloads.
  • Track production bad cases, drive root-cause analysis, and close the loop back into the evaluation framework to systematically prevent regressions.
Requirements
  • 3–5 years in test development, including at least one automation framework built entirely from scratch.
  • Expert-level Python — able to design test frameworks and tooling independently, not just operate off-the-shelf tools.
  • Familiar with API testing, load testing (Locust / JMeter), and CI/CD integration (GitLab CI / Jenkins).
  • Understand the fundamentals and common failure modes of LLM / RAG systems (hallucination, retrieval misses, incorrect tool calls).
  • A considered methodology for testing non-deterministic systems — comfortable with quality judgments that have no single correct answer.
  • Strong data instincts; able to describe quality in metrics rather than impressions and deliver standalone quality analysis reports.
Nice to Have
Ragas / DeepEval / Promptfoo AI safety / red-teaming experience SaaS / financial services QA Familiarity with Java stack
📧 請將簡歷發送至：[email protected] 📧 Send your resume to: [email protected] 立即申請 → Apply Now →
聯絡我們Get in Touch

開啟您的
AI 之旅 Start Your
AI Journey

無論您正在探索想法、評估方案，還是準備落地實施，我們的專家團隊隨時準備提供專業建議，助您在 AI 時代搶佔先機。 Whether you're exploring ideas, evaluating solutions, or ready to implement — our expert team is here to provide professional guidance at every stage of your AI journey.

📧
電子郵件Email [email protected]
📍
辦公地址Address Level 18 Centennial Tower, 3 Temasek Avenue Singapore 039190Level 18 Centennial Tower, 3 Temasek Avenue Singapore 039190
☎️
電話Phone +65 80649006+65 80649006
💬
即時諮詢Live Chat 週一至週五 9:00–18:00Mon–Fri 9:00–18:00

發送訊息Send a Message