全栈机器学习系统工程师
Full-Stack ML Systems Engineer
关于我们
我们是一家以研究为驱动的精英人工智能基础设施初创公司，正在构建企业多智能体应用的定义性工作流智能和策略管理层。虽然市场充斥着用于构建简单AI代理的工具，但没有人成功解决在生产环境中高效、安全且成本效益地扩展这些代理所需的基础设施缺口。我们通过将深度系统编程、软件定义网络和操作系统级原语应用于AI工作流编排来填补这一缺口。
我们最近完成了500万美元的种子轮融资，并作为一个扁平化、高度技术化的三人团队运营。我们的领导团队包括：
- 首席执行官：一位三次深科技创始人（博士，前贝尔实验室，前高通），成功将其上一个项目从种子轮扩展到B轮融资。
- 联合创始人：德克萨斯大学奥斯汀分校的州立主席教授，世界知名的机器学习系统研究员，其背景涵盖卡内基梅隆大学和斯坦福大学。
我们的核心技术源自顶尖学术研究小组。这不是另一个API封装——这是一项为企业的下一个前沿领域设计的基础性基础设施方案。
你将负责的工作
- 连接研究与生产：直接与我们的首席架构师和CEO合作，将尖端的机器学习系统研究转化为产品开发者喜爱使用的企业平台。
- 构建编排基础架构：设计并实现全栈机器学习系统，用于多智能体工作流编排，融合深厚的后端系统能力与简洁、愉悦的开发者接口。
- 构建扩展层：开发由策略驱动的扩展机制，管理多智能体交互，解决限制智能体在大规模下自主性的基础设施瓶颈。
- 主导开发者体验：将研究级别的想法转化为可读性强的内容，通过撰写技术文档、代码示例、示例应用和直观的入门流程来实现。
- 系统集成：跨分布式系统设计、网络配置和异步事件流进行协作，直接应用于运行时AI操作。
- 我们寻找的人选
- 经验与职级
- 5年以上机器学习系统工程的生产经验，或来自顶级机构的相关机器学习系统/分布式计算领域的博士学位。
- 核心技术能力（关键）
- 实战型智能体框架：有使用LangGraph、LangChain等框架将多智能体应用部署到生产环境的实操经验
查看英文原文
About Us
We are an elite, research-backed AI infrastructure startup building the defining workflow intelligence and policy management layer for enterprise multi-agent applications. While the market has flooded with tools to build simple AI agents, nobody has cracked the infrastructure gap required to scale them efficiently, safely, and cost-effectively in production. We bridge this exact gap by applying deep systems programming, software-defined networking, and OS-level primitives to AI workflow orchestration.
We recently closed a $5M Seed round and operate as a flat, hyper-technical team of three. Our leadership includes:
- The CEO: A 3x deep-tech founder (PhD, ex-Bell Labs, ex-Qualcomm) who successfully scaled his last venture from Seed to Series B.
- The Co-Founder: A Regents Chair Professor at UT Austin and world-renowned ML systems researcher with a pedigree spanning CMU and Stanford.
Our core technology originates from a top-tier academic research group. This is not another API wrapper—this is a foundational infrastructure play designed for the next frontier of enterprise software.
What You'll Be Doing
- Bridge Research & Production: Work directly alongside our Chief Architect and CEO to translate cutting-edge ML systems research into an enterprise platform that product developers love to use.
- Build the Orchestration Fabric: Design and implement full-stack ML systems for multi-agent workflow orchestration, merging deep backend systems capability with clean, satisfying developer interfaces.
- Architect the Scaling Layer: Develop the policy-driven scaling mechanics that govern multi-agent interactions, resolving the infrastructure bottlenecks that limit agentic autonomy at scale.
- Own the Developer Experience: Take research-grade ideas and make them legible by authoring technical documentation, code examples, sample applications, and intuitive onboarding flows.
- Systems Integration: Collaborate across distributed systems design, networking configurations, and asynchronous event streams applied directly to runtime AI operations.
What We're Looking For
Experience & Seniority
·
5+ years of production experience engineering ML systems, OR a PhD from a top-tier institution in a relevant ML Systems/Distributed Computing field.
Core Technical Competencies (Critical)
- Hands-on Agentic Frameworks: Proven experience building and deploying multi-agent applications to live production environments using frameworks like LangGraph, LangChain, CrewAI, AutoGen, Semantic Kernel, ADK, or custom orchestration systems.
- Production Scaling & Compute: Direct exposure handling scale on either the orchestration side (managing complex state, cyclical routing, and memory loops) or the infrastructure side (scaling AI compute fabrics, asynchronous jobs, queues, and container clusters).
- Model-Serving Platforms: Practical production experience deploying and optimizing models via frameworks such as vLLM, SGLang, Ray, NVIDIA Triton, or NVIDIA Dynamo.
- Full-Stack Implementation: Strong Python background combined with full-stack capabilities, including constructing front-end visual telemetry dashboards (using Grafana, Tableau, Streamlit, or similar).
- Open Source Focus: Experience utilizing, developing upon, or contributing directly to open-source software repositories.
Mindset & Soft Skills
- Enterprise Product Instincts: The ability to look at powerful backend capabilities and figure out how to make them readable, structured, and satisfying for enterprise end-users.
- Thriving in Ambiguity: An early-stage startup mentality—comfortable shipping fast, iterative code within dynamic, shifting, and initially ill-defined parameters.
- Clear Communication: Strong written skills to help transform complex technical primitives into clear developer documentation and clean product language.
- A background as a Solutions Architect or Forward Deployed Engineer is a significant plus.
Why You Should Join Us
- Solve an Unsolved Problem: Nobody has cracked the gap between building a toy agent app and scaling it across a global enterprise. You will build technology that fundamentally does not exist anywhere else.
- Unmatched Pedigree: Learn from and build with a world-class founding team combining serial entrepreneurship with elite academic systems research.
- Founding-Level Impact: As Employee #4, your code dictates the core blueprint of the platform. You get a direct up-to-1% equity block in a venture backed by a highly technical $5M validation check.
Originally posted on Himalayas