Workflow / AI Engineer

The 90-second pitch

We're a Houston-based luxury lighting and luxury home automation manufacturers rep agency covering a four-state region — Texas, Oklahoma, Arkansas, and Louisiana. We rep premium brands into custom-integration dealers, design firms, and builders across the territory.

We're on a hard push to automate everything that doesn't need a human touching it — so our humans can spend more time on the work that does. Our automation stack already runs Hermes Agent harness, n8n, LiteLLM for model routing, AgentMail MCP, Ollama with a local Qwen model, plus cloud-routed Anthropic Claude and OpenAI GPT for the heavy reasoning. Data lives in Supabase (Postgres + pgvector). We're heavy into Salesforce (with the Salesforce MCP), Asana, and Slack.

We need someone in Colombia or Costa Rica to come in as a long-term partner, pair daily with our Operations Manager and General Manager to identify workflows worth automating, and turn them into agents that actually run. You'll also get exposure across the whole organization — sales engineers, lighting design, marketing, inside sales, project management, all the department heads. That cross-org reach is part of why this role is interesting: you'll see how every part of a real business actually works, not just one slice.

This is NOT a software-engineering-for-the-sake-of-engineering role. It's an operations engineering role where the operations happen to be wired with LLMs, agents, APIs, and integrations. If you love taking a manual process the team hates, mapping it, and replacing it with an agent that runs while everyone's asleep — keep reading.

A note on the culture: We're focused on results, not tenure. We move fast, we let people work on the edges of what they know, and we measure people by what they ship. If that's the kind of place you've been looking for, you're going to fit.

What you'd actually do

You'll touch every part of the business:

Outside sales workflows — dealer prep, follow-up cadence, opportunity scoring, CRM hygiene automation across the territory.
Inside sales automation — lead routing, dealer touchpoint scheduling, quote-to-order tracking.
Marketing automation — email cadences, social posting workflows, dealer co-op marketing programs, content distribution.
Lighting design operations — RFP intake, specification request routing, project-status tracking, design-handoff workflows.
Project management — Asana orchestration, cross-team notifications, deadline surfacing.
Chatbots and training — building product-knowledge chatbots for our dealers and internal training assistants for the team.
Salesforce + MCP integration — wiring Salesforce data into agentic flows, automating common SF operations.
Pair daily with the Operations Manager and General Manager to map a process, propose the automation, ship it, measure it, iterate. Work alongside every other department head — sales engineers, lighting design, marketing, inside sales, project management. The work is collaborative, not heads-down-headphones-on.
Own the BMCIS LATAM recruiting machine as one of your first projects — the same one that hired YOU is the one you'll keep building. Meta, we know. Cool, right?
Document everything you build so the team (most of whom are not engineers) can run, debug, and request changes.

You're a fit if you have

Must-haves

You're a real communicator. This is the most important thing on the list. You're working across every department head in the company — operations, sales, marketing, lighting design, project management. The technical work is teachable; the ability to ask the right questions, surface issues fast, push back when something doesn't make sense, and keep people in the loop without being prompted — that's the part we cannot teach. If you tend to go silent when work gets hard, this isn't the role. We want someone who fires off "hey, I'm stuck on X, here's what I've tried, what do you think?" within 30 minutes, not three days.
n8n in production — you've built and maintained workflows of meaningful complexity (multi-trigger, branching, error handling, custom code nodes), not just hello-world tutorials.
API wiring — you can read OpenAPI/REST docs, authenticate with OAuth, handle pagination, deal with rate limits, and clean up messy webhook payloads.
Light JavaScript or Python — comfortable enough to write a code node when n8n's built-in nodes don't get you all the way there. We don't need a 10-year dev. More important than writing code is being able to read what Claude or GPT writes and check it. You'll be working with AI code assistants every day, and we need you to catch when they're wrong before you ship it.
LLM API exposure — you've made calls to at least one of: Claude (Anthropic), GPT (OpenAI), Gemini, or an open-source model via Ollama, LM Studio, or LiteLLM. Bonus: you've built something with structured outputs / tool calling.
Functional English — you can read technical docs and write clear async messages. Not customer-facing, so accent/fluency don't matter. We'll confirm in conversation; no formal test.
Valid US tourist visa (B1/B2) — already issued. We bring the team to Texas twice a year for in-person events. The visa process can take a long time depending on country, so we need you to already hold a valid B1/B2. We verify the visa with you live during the Stage 1 interview.
Schedule overlap — daily overlap with our Houston team needed (US Central Time). We're flexible on exact hours within your day, but expect to be available for collaborative work during that window.
Reliable home setup — quiet workspace, decent internet, good headset/camera for daily team calls.

Nice-to-haves (any of these moves you up the stack)

Direct experience with an agentic harness (Hermes, LangGraph, CrewAI, AutoGen, or similar).
LiteLLM, OpenRouter, or model-routing experience.
Vapi or other voice-AI platform experience.
Salesforce experience — REST API, Apex, or Salesforce MCP (huge plus, even partial exposure).
Supabase / Postgres with pgvector experience.
Asana API or automation experience.
Slack agent / app configuration — building bots that operate in channels, slash commands, interactive messages.
An active GitHub profile with real work — even small repos. Helps us see how you think.
Experience with vector databases and RAG patterns.
Comfortable with Docker and OrbStack for managing local services.

You're probably NOT a fit if

You want to be a "real software developer" building products from scratch. This is automation engineering. Different muscle.
You hate documentation. Half the value of this role is leaving behind things others can run.
You need 100% async work. We sync daily with the ops team and live in Slack.

Compensation, terms, and what we offer

Compensation: USD $2,200/month, paid in USDC stablecoin to a wallet of your choosing. The rate isn't strictly negotiable, but for the right person who proves themselves in the interview, we'll go higher. Show us in the conversation.
Bonuses and raises. We treat our contractors like team members, not vendors. You'll be in our bonus program within your first year. Your first bonus comes at the end of year one, and we review and raise the monthly rate on a periodic schedule based on performance and scope. We're building a long-term bench — we want you here for years, not months.
Engagement type: Independent contractor (US-LATAM), twice-monthly invoicing paid on the 15th and the last business day of each month, 14-day termination-for-convenience either side.
Hours: Full-time-equivalent (~160 hours/month), with daily Houston overlap.
Equipment: You provide your own computer and internet. We may be able to help with office setup costs if needed — talk to us.
Travel: 2 paid trips to Texas per year for company events; at least one annual team event in your region.
Tech stack growth: The major pieces are set, but we're open to growing the stack further if you make a reasonable case.
What we DON'T provide: No health insurance, no local pension or social-insurance contributions, no PTO accrual. You're an independent contractor; this is built into the rate.
What we DO offer: A serious tech stack, a team that respects engineering autonomy, a direct line to the GM (James), and the chance to build agentic AI infrastructure that's actually in production — not a toy demo.

Our culture (read this part — it matters)

Work hard. Up early. Deliverables done. Go the extra mile. We don't sugar-coat, we surface issues fast, and we help each other when something's stuck. Slack is our nervous system.
Results, not tenure. We measure people by what they ship, not how long they've been around. Move fast. Learn on the edges of what you already know. Take swings.
Family, not clock-in, clock-out. People stay with us a long time because we run it like a tight ship with a warm crew. Everyone helps everyone. New hires get pulled in fast.
In-person matters. When you visit Texas twice a year (or when we come to visit your team in their region), we work hard during the day but we have a real good time in the evenings together.
The corporate flow is dead to us. If you're tired of meetings about meetings, ten layers of approval, and pretending to care about things that don't matter — this is the way out.

Timeline

We're ready to move fast: application to offer typically 14-21 days.

Referrals

The role above is one of four we're hiring directly into our team over the next quarter. If you know other strong people in Colombia or Costa Rica who'd be great here (we'll need a marketing manager, a senior Next.js developer, and a marketing account manager next), refer them and we'll honor a referral bonus when they're hired.