Founding Machine Learning Engineer

arcadeaiSan Francisco, CAEngineering

Apply on company site

Opens the official arcadeai careers page

Salary

Not listed

Job type

Full-time

Experience

Lead · 7+ yrs

Posted

5 Oct 2026

Status

Open · checked 7d ago

About this role

This role suits an experienced machine learning engineer with a strong background in training and shipping production models who wants to build enterprise AI agent infrastructure in San Francisco, CA. The position involves owning the model pipeline from data ingestion to publishing for agent capabilities.

Role Overview

Arcade is hiring a Principal Machine Learning Engineer in San Francisco, CA to build the future of agent capabilities. This role involves designing and owning the machine learning models that sit in the runtime path of every agent call, focusing on enterprise readiness, small custom models, and production-grade deployment.

Key Responsibilities

Own Arcade's ML pipeline end to end, covering data, training, evaluation, and release processes.

Fine-tune and train models for tool selection, routing, retrieval, and agent memory using distillation, embeddings, and rerankers.

Expand model use cases into search and recommendation over tools and agent context.

Build offline and online evaluation systems to compare models head-to-head with frontier models using real agent traces.

Quantize, optimize, and package models for customer VPCs and air-gapped environments.

Turn production agent traces and tool-call data into training data while respecting enterprise data boundaries.

Set the ML stack strategy, choose tools, and guide future hiring in machine learning.

Decide which features to build using depth in modern agent systems like harnesses, memory, and skills.

Requirements

7+ years of software engineering experience.

4+ years training and shipping production machine learning systems.

Experience training or fine-tuned models that reached production and impacted customer-facing metrics.

Expertise in production agent systems including harnesses, memory, skills, tool use, and sub-agents.

Practical knowledge of fine-tuning applications, required data types, and deriving data from telemetry.

Statistical fluency to evaluate model performance and differentiate real deltas from noise.

Experience deploying models under constraints such as tight latency budgets, limited GPUs, or external infrastructure.

Skills and Tools

vLLM

ONNX

TensorRT

Embeddings

Rerankers

Classification models

Recognition models

Fine-tuning frameworks

About the Company

Arcade provides an MCP runtime that connects AI agents to enterprise systems with permission slips and audit trails. The company is backed by SYN Ventures, Morgan Stanley, and Wipro, and serves Fortune 100 enterprise customers like Morgan Stanley and Open Table.

How to Apply

Apply on the official company careers page.

Skills and tech stack

vLLM ONNX TensorRT LangChain LlamaIndex Python LLM TypeScript Machine Learning Fine-tuning Distributed Systems

Questions candidates ask

Where is this job located?

This position is located in San Francisco, CA.

What experience is required for this role?

Candidates need 7+ years of software engineering experience, with 4+ years focused on training and shipping production machine learning systems.

Who will the role report to?

The role reports directly to the Head of Engineering.

What kind of environments do the models need to run in?

Models must be designed to run in customer VPCs, customer hardware, and air-gapped networks with strict latency and infrastructure constraints.

What is the primary focus regarding model development?

The focus is on small, fast, accurate, and custom models that can be shipped directly into enterprise customer environments for agent tool selection and routing.

Ready to apply?

The application is on the official arcadeai careers page.

Apply on company site