Platform Engineer - Inference

About the role

Subconscious is building the token factory for long-running agents. Long-running agent use cases for things like coding, deep research, and computer-use already consume the majority of inference capacity today, and we expect the demand for this type of inference to grow exponentially.

Our core technology enables us to serve these context-heavy workloads (100+ LLM requests, 200k+ context) with frontier efficiency and performance. You will provide the battle-tested infrastructure experience needed to scale the platform to thousands of users. Today, we serve billions of tokens per day. With your contributions, we will serve trillions.

This is a ground-floor role. You will shape the product, culture, and technology. We expect the person who fills this role to grow into a leadership position.

Responsibilities

• Own the reliability, performance, observability, uptime, load balancing, and security of inference workloads.
• Work closely with our research team to make new inference optimizations and open source model support broadly available to customers.
• Contribute to the technical direction of our inference platform including major architecture and build vs. buy decisions.
• Help build the product factory that makes our team ship code like a team 10x our size - including human-driven process, CI/CD, and agent guardrails.
• Establish durable engineering culture through design docs, architecture reviews, mentorship, and hands-on implementation.

You will work across our inference stack, from the user-facing platform to GPU cache optimization.

Requirements

• Hard skills:
  • 5+ years in Platform, Infrastructure, or Inference engineering roles
  • Experience running Kubernetes at scale
  • Experience running AWS at scale
  • Deep observability experience - including Prometheus, Grafana, and Datadog
  • Expert in deployment jujitsu - zero downtime, rollbacks, blue/green, canary, etc. - we want someone with battle scars
  • Proficient in Terraform, CI/CD pipelines, and GitOps workflows
  • Proficient in Python, TypeScript, and/or Rust
  • Experience with SDLC security and compliance processes
  • Experience with agentic software development
• Soft skills:
  • Staff-level judgment - you can define ambiguous problems, make pragmatic tradeoffs, influence without authority, and leave both systems and teams better than you found them. You can tell a good story with data.
  • Product sense - you solve problems by working backwards from the end-user and business requirements.
  • You have character, a sense of humor, and compassion. You are someone we are excited to introduce to our partner.

Nice to have

• You have worked on systems that serve trillions of tokens per day
• You have worked in startups before
• Experience with vLLM, SGLang, or NVIDIA Dynamo
• Experience with BYOC (bring your own cloud) implementations
• Experience with SOC 2 or FedRAMP - ideally having led or contributed to achieving third-party attested compliance
• Experience with GCP or Azure
• Experience with agent-driven software delivery at scale

Location

We are based in Boston, MA, and are in the office 5 days a week. For an exceptional candidate, we are willing to consider remote work.

Compensation

We can't build this company without you, and we want you to be a meaningful owner alongside us. Your equity will reflect that.

Salary will be between $200k-$250k with significant equity ownership.

Benefits

We offer full medical, dental, and vision coverage, plus unlimited PTO.