This Software Developer is a core builder on a brand-new agentic infrastructure company funded by Martell Ventures — one of the first engineers shaping how agents are executed, orchestrated, and scaled. You'll work across a Go-heavy backend, Next.js frontend, and bare-metal/AWS infrastructure to ship the primitives customers depend on, and you'll do it at a pace that matches a well-funded zero-to-one sprint. The bar isn't code volume — it's genuine AI fluency: you understand how LLMs work at the model level, you've built real tools with them, and you can make principled decisions about cost, latency, and capability tradeoffs without being hand-held.

Responsibilities

Build and extend the core agent execution and orchestration platform in Go
Integrate and manage LLM providers, handling model selection, cost management, and API reliability across the stack
Containerize and deploy services using Docker across bare-metal and AWS environments
Contribute to the Next.js frontend surfaces that expose agent workflows and platform controls to customers
Design and maintain Postgres schemas and queries that support durable agent state and memory
Instrument and operate the platform across deployment environments, tracing failures from infrastructure up through agent behavior
Move fast on new capabilities — scope is tight, iteration cycles are short, and priorities shift as customers do
Participate in architecture conversations and bring informed opinions on tooling choices, not just implementation

Must-haves

Production experience writing Go (backend services, APIs, or systems tooling)
Demonstrated history of building software that uses LLMs — shipped projects you can walk through, including decisions around model cost, context management, and failure modes
Hands-on comfort with Docker and AWS (deployment, debugging, and operations — not just config files)
Ability to work in a zero-to-one environment: minimal process, high ownership, fast context-switching
Genuine understanding of how LLMs work — tokenization, context windows, cost curves, model tradeoffs — not just API call familiarity

Nice-to-haves

Experience with agent frameworks or orchestration patterns (tool use, multi-step planning, memory retrieval)
Familiarity with bare-metal infrastructure provisioning and networking
Prior startup or venture-studio experience where you shipped before the team was fully formed

How the hiring process works

01
Apply through the button above. Martell Ventures reviews your application.
02
If there's a fit, Anvil sends you a private take-home tailored to this role. Three to five focused hours, plus a short Loom walk-through.
03
Your submission is graded against a known rubric. You hear back with real feedback. If both sides want to continue, you move to a conversation with the team.