Machine Learning Engineer

Sweep360•New York, NY

3d•Onsite

About The Position

We are building humanity's defense layer for the AI age and are looking for an exceptional machine learning engineer to build the AI decision system that turns raw signals into trusted operational decisions across device, cloud, and offline environments. This role is akin to joining early Tesla to make Autopilot work in the real world or Figure AI to build the firmware that made them trustworthy. Sweep is deploying alongside the world’s highest-stakes teams, including Olympic delegations, F1 paddocks, and senior government officials, to ensure the intelligent machines we rely on remain aligned with us. We are a small, talent-dense team with high ownership, high velocity, and low ego, focused on building something that outlasts us and redefining cyber-physical security for the AI age. This is the first dedicated AI systems hire, and you will be the difference between a system that exists and one that works, ensuring the reliability of the entire AI system from data ingestion to operator decision. You will turn noisy cyber-physical observations into trusted operational decisions, define how the system reasons under uncertainty, and your work will be used in high-stakes environments where outputs must be trusted. You will become the technical lead for Sweep’s AI decision system before Series A.

Requirements

5–10 years owning production systems end-to-end.
Strong system design across APIs, pipelines, and data storage.
Built production AI systems trusted in real-world operations.
Strong Python, plus Go/TypeScript (or similar).
Comfortable building systems spanning edge devices, cloud, and intermittent connectivity.
Able to debug production systems quickly and decisively.
Communicates clearly and operates independently.
U.S. Person status required (may involve export-controlled data).

Nice To Haves

Handled streaming systems (Kafka, pub/sub).
Created production LLM or inference pipelines (prompting, retrieval, evaluation).
Designed for adversarial or security environments.
Built systems that run on-device as well as in the cloud.
Thrived in an early-stage startup environment.

Responsibilities

Shape how the production AI system behaves in the real world.
Design ingestion → reasoning → decision systems.
Drive inference reliability, predictable logic, and transparent reasoning.
Close the loop from deployments → system learning.
Make the system trusted under real-world conditions.
Partner with RF / hardware / field teams to deliver for elite users globally (~10–15% travel).