Principal Engineer - The Starr Conspiracy - NORTHAMPTON

Principal Engineer

The Starr Conspiracy

Northampton, MA (In-person / Hybrid)

Hello. My name is Bret Starr. I am the founder of The Starr Conspiracy. I’m looking for a Principal Engineer.

TL;DR

We're a 10-person B2B marketing agency with a 25-year legacy. But in 2022, I saw an asteroid headed towards planet professional services, so I made the difficult decision to pivot away from hours-based agency work and towards AI-native solutions for B2B marketing.

Today, we're nearly four years deep into an AI-native transformation. It's not been easy. We've been building great stuff for at least a couple of years now, but no one understood (much less wanted to buy) AI-native marketing solutions. Well, that was until the beginning of this year.

Now we have the opposite problem. (The proverbial "good problem to have.") We're struggling to keep up with demand for our products.

By the way … we still provide classic B2B consulting services. We believe the combination of our world-class marketing services (branding, messaging, GTM strategy) and AI-native solutions is what makes us unique and valuable. But our primary goal is to transition from majority services revenue to majority tech revenue. And we're already about halfway there.

We have shipped several bootstrapped, AI-native products currently used (and loved) by more than a dozen B2B tech companies. And demand for our solutions grows every week. Our products actually work (which is more than you can say for most AI stuff these days … especially marketing solutions). And our plan is to parlay our tech (and our tech revenue) into an acquisition within three to five years. And when we are acquired, our valuation will be as a technology company, not a services company.

That's the plan anyway. Always easier said than done, right? And who knows what the world will throw at us next.

Anyway, here's the upshot. We need a senior engineer who can do two things at once: ship code alongside our existing dev team, and build the architectural backbone that turns a stack of really good client solutions into a cohesive platform. In short, we need a little help cleaning up our act.

Why this role exists

Our products work. Our dev team hustles. And our clients get results. What we don't have is architectural discipline, coherence across the portfolio, documented systems a stranger could extend, or a technical narrative an acquirer can follow. Right now, a lot of that knowledge lives in the founder's head (that's me). That's fine for right now. But when it's time to exit, that will kill our valuation.

We're looking for the person who fixes that. You will lead the architecture conversation, ship the platform layer, mentor the devs, and be the technical voice that a diligence team talks to when the time comes.

What you'll actually do

About half your time will be hands-on. Building, reviewing, refactoring. You're in the code with our devs, not above us. You bring architectural discipline we don't yet have, the ability to look at five different client solutions, see how they fit together, and build a plan to integrate them in elegant ways on a common standard.

The other half is enterprise value creation. Documenting the architecture. Shaping the technical roadmap. Extracting reusable IP from client work. Occasionally sitting in a client conversation where technical credibility matters. Leveling up our developers and helping us hire more. Building the technical narrative that the company will need at exit.

What this role is not

• Not a CTO. That title would oversell what this is and attract the wrong people. We want a Principal Engineer. Someone in the trenches. Not someone on a hill.
• Not a manager. We have a small (2 people), dedicated dev team (and that includes me). We also receive contributions from others in the company. That's it. That's the team. So if you need twenty reports to feel important, this isn't for you.
• Not an AI casualty. We've been shipping AI-native systems since 2023. We need someone who's done the same, not someone who caught the wave last quarter. It's been my experience that the biggest blockers to AI transformation inside product companies are the developers themselves. So if you're one of those engineers who got displaced by AI because you wouldn't get on board, and now you're rushing to catch up, this isn't for you.
• Not remote. See below.

Who we're looking for

You need to have spent at least 10 years as a software engineer. And you need to have spent the last two or three shipping real AI-native products. Not prototypes. Real products. You understand what it means to architect software where AI models are first-class system components, where prompts and retrieval and model selection and validation are all engineering concerns.

You can explain complex systems to non-technical people. This matters because our employees, clients, and eventually our acquirers, all need to understand what the founder and dev team have built. Most senior engineers are bad at this. If you're good at it, you're exactly who we want.

You have a small-company temperament. You like being in the work, not adjacent to it. Ambiguity doesn't stress you out. You'd rather ship something imperfect and improve it than design a perfect thing that never ships. You've either worked at a small company before or you're actively trying to escape a big one. Let me be explicit. Our company is terrible at training people. It's a sink or swim environment with a lot of autonomy. The founder is a perfectionist who iterates too much before shipping. People succeed or fail based on their own ability to hold themselves accountable and to a high standard.

Where this is

In person, at our Northampton, Massachusetts office (and a mighty fine office it is, if I do say so myself). This is a choking-distance role for reasons that matter to how architectural decisions get made. You and the founder need to be in the same room often enough that the thinking compounds.

If you want to work a couple days a week from home, that's fine.

We're not looking for commuters from Boston. We're looking for someone who lives in (and loves) Pioneer Valley. And we plan to do a significant amount of our future hiring right here in the Valley. (We have a satellite office in Fort Worth, and will be hiring there as well, JSYK.)

Tech StuffLanguages & Frameworks

• TypeScript + Next.js (App Router, currently on 14, planned upgrade to 16): most of the user-visible product
• Python 3.11+ + FastAPI: the entire pipeline backend
• React 18 + React Flow (@xyflow/react): the visual pipeline builder

Data Layer

• MongoDB Atlas: not just CRUD; our products lean heavily on aggregation pipelines, partial unique indexes, atomic findOneAndUpdate for concurrency claims, and TTL collections
• Multi-tenant scoping (client_id + analysis_run_id on every read/write) is invariant-level: this is the load-bearing pattern for the entire data architecture

Hosting & Infra

• Vercel: Next.js dashboard, dev + prod as separate projects
• Railway: Python pipeline + FastAPI service
• GitHub Actions: CI with Pytest + next build + npm audit + Vitest hard gates

Auth & Identity

• Clerk: organization-based multi-tenancy; product entitlements live in publicMetadata, not in our DB

LLM / AI Stack

• Anthropic Claude API (Opus / Sonnet / Haiku)
• OpenAI API (GPT-5+)
• Comfort with server-side LLM output validation

Domain Stack (learnable, but candidates should know it exists)

• AEO (Answer Engine Optimization): schema.org structured data, multi-engine citation tracking across Perplexity / ChatGPT / Claude / Gemini / Grok / Google AI Overview
• SEO data harvesting: SerpAPI, DataForSEO, AlsoAsked, Ahrefs
• GTM concepts: ICP, JTBD, demand-state modeling, primer scoring
• Content and creative generation
• Deep semantic analysis

Engineering Patterns That Define the Codebase

• Cron concurrency hardening: CRON_ROLE env gates, atomic claim collections, partial unique indexes (we got bitten by this; it's now invariant)
• Feature-flag cutovers: NEXT_PUBLIC_* flags for graduated rollouts; legacy paths preserved byte-for-byte until canary passes
• Cross-project MCP integration: the GTM Kernel (more about that when we talk) is a separate repo exposed via MCP + REST; we consume it, never write back
• Embeddings + clustering tradeoffs: HDBSCAN vs supervised classification (we learned the hard way that B2B corpora are taxonomic, not density-clustered)

The dealbreaker fundamentals are MongoDB-at-scale, Next.js + TypeScript, Python, and concurrency-correctness instincts. Everything else (Clerk, Vercel, Railway, the LLM APIs, the AEO domain) is learnable in the first month if the candidate has the fundamentals and judgment.

Comp

• Base: $175-225K, depending on experience (and negotiable for the right candidate)
• Exit participation: Meaningful. We're hiring you to build enterprise value and you need to participate in the upside you create. This will not be a token grant.
• Benefits: Full health, dental, vision, 401(k), unlimited PTO … all the good stuff.

How to apply

If this reaches you through someone who knows us, just reply to them. We'd rather talk to warm intros than read cold resumes.

For a more formal path, you can email our COO, Dan McCarron. His email is dan@thestarrconspiracy.com

The Starr Conspiracy is a services-enabled product company in the B2B tech vertical. We build AI-native GTM software and help our clients implement and use them to maximum benefit. We don't hire juniors. We don't do hourly billing. We hire two kinds of people: senior consultants and developers. And they do all the work that matters. Because we believe that's what our clients want, need and deserve: to work with senior people all the time, and access to AI-native solutions adapted to their unique needs.

You need to have spent at least 10 years as a software engineer. And you need to have spent the last two or three shipping real AI-native products. Not prototypes. Real products. You understand what it means to architect software where AI models are first-class system components, where prompts and retrieval and model selection and validation are all engineering concerns.

You can explain complex systems to non-technical people. This matters because our employees, clients, and eventually our acquirers, all need to understand what the founder and dev team have built. Most senior engineers are bad at this. If you're good at it, you're exactly who we want.

You have a small-company temperament. You like being in the work, not adjacent to it. Ambiguity doesn't stress you out. You'd rather ship something imperfect and improve it than design a perfect thing that never ships. You've either worked at a small company before or you're actively trying to escape a big one. Let me be explicit. Our company is terrible at training people. It's a sink or swim environment with a lot of autonomy. The founder is a perfectionist who iterates too much before shipping. People succeed or fail based on their own ability to hold themselves accountable and to a high standard.