What you'll do

Build AI agents that automate repetitive work we're currently doing manually.

You'll create agents for things like: automating document processing, refactoring code, deployment checks, data extraction, and whatever else we're wasting human time on.

You'll also deploy and optimize our own LLMs making them faster, cheaper, and easier to run.

The goal: Turn a 2-hour manual task into a 2-minute automated one. And make LLMs that give answers in seconds, not hours.

Day-to-day work

Automation:

• Build agents that handle repetitive manual tasks for internal teams
• Create workflows that eliminate boring busywork
• Make agents that handle internal processes and requests
• Build code refactoring and review automation
• Create deployment pipeline agents that catch issues early

LLM deployment & optimization:

• Deploy and run self-hosted LLM instances
• Make deployments simpler and faster
• Optimize inference speed: turn multi-hour tasks into minutes
• Reduce costs by making models more efficient
• Profile and fix bottlenecks in LLM pipelines

Integration work:

• Connect our self-hosted LLMs to existing systems and databases
• Make agents that actually work in production, not just demos
• Monitor and improve agent performance over time

You'll probably like this if

You enjoy:

• Building things that eliminate boring manual work
• Making AI actually useful, not just impressive
• Shipping working prototypes quickly
• Deploying and running your own models instead of just calling APIs
• Making LLMs faster and cheaper to run
• Connecting different systems together

You have:

• At least a 3-5 of software engineering experience
• Experience working with LLMs (self-hosted)
• Deployed or run models yourself (not just API calls)
• Built something with AI that people actually use
• Comfortable with Python
• Basic Linux and infrastructure knowledge
• Optimized LLM performance or inference speed

Extra credit if:

• You've deployed open-source LLMs (Llama, Mistral, etc.)
• You've optimized model inference (quantization, batching, caching)
• You've built agents that handle multi-step workflows
• You understand when to use AI vs when a simple script is better
• You've worked with GPUs or model serving infrastructure

This probably isn't for you if

• You only want to use paid APIs, not deploy models
• You spend more time reading about AI than building with it
• You're not comfortable with infrastructure and deployments
• You need everything perfect before releasing

Tech you should be comfortable with

Must know:

• Python (this is what we use)
• LLM deployment and inference (self-hosted models, not just APIs)
• Basic Linux and infrastructure
• Git and basic deployment

Nice to have:

• Agent frameworks (OpenClaw, LangChain, AutoGPT, or similar)
• Model optimization techniques (quantization, vLLM, TensorRT)
• GPU infrastructure and serving
• Docker, basic DevOps
• PHP, SQL, shell scripting

 

What you get

• Real impact: Your agents directly save people hours of manual work
• Cutting-edge work: Deploy and optimize the latest LLMs
• Fast feedback: See your automations working (or breaking) immediately
• Good tools: Latest tech, alpha/beta access, GPUs, decent hardware
• Room to grow: Tech lead, deep specialist, or start something—up to you
• Global team: Smart people from all over + you'll work directly with founders/executives
• Freedom: Own your decisions, move fast, try new things