Python Automation Engineer - AI Integrations

Python Automation Engineer – AI Integrations

2 months ago

Remote

Full Time

Intermediate or Experienced

About BOLD Business

About the job

Automation Engineer — AI Agent Systems

Remote, Latin America | 3-4+ years of actually shipping things
Reports to: Senior Operational Leader

About Bold Business

We build multi-agent AI systems that do real Sales and Recruiting work, not demo-day theater. If it cannot survive a Tuesday with real users, it does not count.

The job

You will build and run agent systems : routing, workers, tools, memory, retries, and the part where we find out it broke at 11pm. You work with a Lead Architect and with Sales and Recruiting humans who will tell you if it is useful. You will not disappear into a cave and emerge with a framework rewrite.

You will also sit with the developers building our CRM, ATS, and ERP. You represent this team in those rooms so those systems can talk to our agents, and our agents can talk back. You are the integration. You are not the CRM owner.

Ask questions. Show the work. If something is on fire, say so the same day. We have already lived the other version.

What you will actually do

• Build Python agents that behave as a system, not as 14 clever prompts in a trench coat
• Wire them to real platforms: APIs, webhooks, data in, data out, no duplicate rows because the webhook got excited
• Work with the teams building our CRM, ATS, and ERP. Show up in their design conversations. Do not leave until the APIs and events actually work for automation
• Debug production like an adult. Trace the bad run to the tool call, the payload, or the prompt
• Log decisions, tool calls, and cost. Alert us before finance does
• Sit with users, watch them use it, then change it

Must have (we will test these, not just admire them on a resume)

Hard

• 3-4+ years of backend, automation, or AI that ran in production. Walk us through something you shipped. "I contributed" is not a walkthrough.
• Production LLM APIs (OpenAI, Anthropic, Groq, or similar). Tool calls, rate limits, token burn, and what a failure looks like when it is not a blog post
• Python someone else can maintain. Tests and CI you actually run
• REST and webhooks: idempotency, retries, timeouts. Pop quiz: it fires twice. What happens.
• SQL past SELECT * . Joins, indexes. Postgres is a plus
• You can read logs and find the root cause. You assume failure is the default and you designed for it

Soft

• Clear English, written and spoken, with a US team. Your application is the first test. Use it.
• You surface problems the same day, not in the postmortem
• You will start without a 40-page spec
• You care if Sales or Recruiting actually used the thing, not just that it compiled
• You can sit with another engineering team, explain what we need in plain English, and not leave until the API and events actually work for automation. Salesforce / HubSpot / Greenhouse / ERP experience is a plus, not a must. Surviving someone else’s system is the must.

Nice to have (we can teach this)

LangChain, LangGraph, LlamaIndex, or whichever library is fashionable this quarter. Vector DBs. AWS/GCP, Docker. Datadog. Prompt evals. If you have them, great. If you don't, and you have shipped systems, we will live.

Why Join Bold Business?

• Build on your Python and integration skills while learning AI systems alongside our senior team.
• Build solutions used directly by Sales and Recruiting teams.
• Work with real APIs, databases, automation, and AI integrations.
• Work remotely with a US-based team with clear collaboration hours and competitive compensation.

Details

Remote, Latin America. Overlap live US hours or this does not work.
Pay is competitive and tied to what you have shipped, not how many libraries you can name. Equipment provided. Health insurance included.

Shipped systems. Clear communication. Curiosity. That is the whole interview.

Remote Work

Preferred Skills

Python (production-level, clean code, automated testing)

LLM APIs & Frameworks (OpenAI, Anthropic, Groq, LangChain, LangGraph, LlamaIndex)

Agent Memory Management (short-term, session-level, long-term)

Tool & Function Calling Integrations

Prompt Engineering & Testing (regression testing, LLM evaluation tools like RAGAS)

System Failover & Reliability (error handling, retries, fallbacks)

LLM Abstraction Layers

Data Pipeline Engineering (ingestion, transformation, validation, quality control)

Cost Optimization & Monitoring (token management, caching, logging, alerting, cost/performance tracking)

AI Security & Privacy (prompt injection protection, least-privilege access, sensitive data masking)

Version Control & DevOps (Git, CI/CD practices, Docker/containerization, AWS/GCP)

High Ownership & Proactivity

Collaborative & Transparent Communication (regular check-ins, progress sharing, early risk reporting)

Strong English Communication (written and verbal for U.S.-based team collaboration)

User-Centric Mindset (empathy for internal teams, workflow understanding, educating non-technical users)

Accountability & Problem-Solving

Requirement Validation & Feedback-Driven Iteration