VP of Engineering

Peach Pilot — Atlanta, GA (Buckhead) / Hybrid

 

We’re not building another AI wrapper. We’re building an AI workforce that earns the right to do real work.

Peach Pilot is a multi-agent autonomous AI platform that does real work — executive briefings, real-time research, psychological profiling, compliance monitoring, and full software development. We have dozens of specialized agents today and the architecture to scale to hundreds.

Our agents don’t wait for prompts. They think, coordinate, and deliver. Built on OpenClaw, our orchestration layer lets us deploy, manage, and scale agent teams that work alongside humans — not as chatbots, but as capable coworkers.

But here’s what we’ve learned: the technology only works if the humans trust it. In regulated financial services, AI that moves fast without earning trust gets shut down. That’s why our architecture separates the agents that do the work from the governance layer that ensures every action is auditable, explainable, and human-supervised. We call this our dual-agent architecture — execution agents handle workflows while governance agents monitor compliance, flag risks, and route exceptions to human decision-makers through Mission Control.

We need the engineering leader who understands that building trust is an architecture problem, not just a product problem.

The Opportunity

This is a foundational hire. You’ll be the first VP of Engineering at Peach Pilot, reporting directly to the CEO. There are no legacy systems to untangle, no bureaucratic layers to navigate, and no ambiguity about impact — the platform you build will be the platform.

Why now: We have a working multi-agent system with dozens of specialized agents, a clear product roadmap through mid-2026, and infrastructure supported by a dedicated team covering infrastructure, IT, and modeling. What we don’t have is the engineering leader who will take this from early platform to scalable product. That’s you.

Our advantage: Peach Pilot is a fully independent, funded startup — not a division or internal project. What gives us an unfair advantage is strategic access to sister companies spanning life insurance, payments processing, loan origination, and other financial systems businesses. Their real data, real systems, real processes, and real people accelerate our path to product-market fit so we can build and validate faster than any startup starting from cold outreach. Once we’ve proven the platform, we sell to the broader market. You’ll be building production AI for regulated financial services from day one, with the domain access that most startups spend years trying to negotiate.

Why this matters: The multi-agent AI space is moving fast, but most companies are building demos. We’re building production systems where accuracy has compliance consequences, where bias detection isn’t optional, and where “hallucination” isn’t a punchline — it’s a regulatory risk. Financial services is a relationship-driven industry where AI can transform how organizations operate, from underwriting support to client management to compliance. That’s exactly why it’s the most interesting.

 

What You’ll Do

Build on the Platform we Started (Months 1–3)

• Own architecture and delivery of a multi-agent orchestration system built on OpenClaw: agent routing, context sharing, cross-agent handoffs, and conflict resolution across dozens of specialized agents
• Design and implement the dual-agent architecture: separate execution agents (that do work) from governance agents (that monitor, audit, and escalate), ensuring every sensitive action is auditable and human-supervised
• Build Mission Control — the real-time governance dashboard where human operators monitor agent activity, review escalations, track SLAs, and maintain oversight of autonomous workflows
• Scale the Organizational Context Graph (Cosmos DB vCore + vector search) — the persistent knowledge graph that captures institutional memory, organizational dynamics, and decision patterns to create compounding intelligence unique to each client
• Design and ship ingestion pipelines processing 500+ news events/day with NER, embedding generation, deduplication (cosine similarity), and multi-source verification
• Integrate and optimize across three LLM providers (Anthropic Claude, OpenAI GPT-4o, xAI Grok-4) with intelligent model routing and fallback
• Ensure 99%+ uptime for systems where accuracy matters — no “move fast and break things” here

Build the Team (Months 1–3)

• Grow engineering from ~3 to 12–15, hiring across backend, AI/ML, data, and infrastructure
• Establish engineering culture from scratch: code review standards, testing practices, deployment pipelines, incident response, on-call rotations
• Define sprint cadences, delivery metrics, and engineering OKRs that actually mean something
• Create the kind of team that attracts people who build things that matter

Build the Strategy (Ongoing)

• Partner directly with the CEO on roadmap prioritization and technical feasibility
• Make foundational architecture decisions: database choices, agent communication protocols, multi-tenant isolation patterns
• Evaluate build-vs-buy across the stack — we’re opinionated about building what differentiates us and leveraging what doesn’t
• Shape the technical vision for how AI agents collaborate at scale in regulated industries — where trust, governance, and audit trails are non-negotiable

What You’ll Build With

The stack below reflects what our sister companies currently run in production. As an independent startup, we have the freedom to architect with the stack that best serves our platform — we’re not locked into inherited decisions. That said, Azure and this tooling give us a strong starting point with real production workloads to build against.

Layer	Stack
Cloud	Azure (Cosmos DB, AI Search, Functions, Service Bus, Blob Storage, Monitor)
AI / LLM	Anthropic Claude (Opus, Sonnet), OpenAI GPT-4o, xAI Grok-4
Data	Cosmos DB vCore + vector search, Azure AI Search (hybrid semantic + keyword)
Pipeline	Azure Functions (Python), RSS/API ingestion, NER/embedding pipelines
Orchestration	OpenClaw-based multi-agent system with cross-agent API communication
Governance	Mission Control dashboard, dual-agent architecture (execution + oversight)
Languages	Python (primary), TypeScript/Node.js
Analytics	Hex.tech, Snowflake

 

Who You Are

You Have

• 10+ years in software engineering with 3+ years leading engineering teams (VP, Director, or Head of Engineering)
• Production AI/ML experience — you’ve shipped models and AI systems that real users depend on, not just notebooks and POCs
• Cloud-native architecture depth on Azure (or strong AWS/GCP with genuine interest in Azure)
• Hands-on experience with vector databases and embeddings (Cosmos DB vector, pgvector, Pinecone, Weaviate, or similar)
• Distributed systems expertise — event-driven architectures, message queues, microservices at scale
• Team-building track record — you’ve grown engineering teams from small (3–5) to mid-size (15–25) and know the cultural inflection points along the way
• Comfort as both architect and engineer — in the early months, you’ll review PRs, debug pipelines, and make schema decisions alongside your team
• An intuition for trust-building in enterprise software — you understand that in regulated industries, how the system earns human confidence is as important as what it can do

 

Even Better If You Have

• Experience in financial services, fintech, or regulated industries where compliance isn’t a checkbox
• Background in NLP/NER pipelines, knowledge graphs, or multi-agent AI systems
• Familiarity with graph databases (Neo4j, Cosmos DB Gremlin) and real-time data processing
• A 0-to-1 track record — you’ve built products from scratch, not just maintained existing ones
• Experience with the Anthropic API, Claude, or Model Context Protocol (MCP)
• Hands-on experience with OpenClaw — our agent orchestration layer runs on it, and familiarity with its architecture, skill system, and session management is a real advantage
• Experience designing human-in-the-loop governance systems, real-time monitoring dashboards, or compliance automation for regulated environments

What We Don’t Require

• A specific degree (we care about what you’ve built, not where you studied)
• Big-company pedigree (startup experience is often more relevant)
• AI research publications (engineering excellence > academic credentials)

 

Compensation & Benefits

• Base salary: $200,000 – $280,000 (commensurate with experience)
• Equity: Meaningful ownership stake in Peach Pilot — this is a foundational leadership hire, and the equity reflects that
• Benefits: Comprehensive health, dental, and vision insurance; 401(k); flexible PTO
• Location: Hybrid in Buckhead (Atlanta, GA) — we value in-person collaboration for the core team, with flexibility built in

 

What Makes This Different

You’ll have a seat at the table, not a seat in the back. You report to the CEO. You’re in the room when strategy is set, not downstream when tickets are written.

The tech is real. Dozens of specialized agents running in production today, with architecture designed to scale to hundreds: Chief of Staff, Research, Insight (psychological profiling), News (real-time world knowledge), SDLC agents, Action Extractor, Execution, and more. Built on OpenClaw, our platform empowers every employee at a client organization to work alongside AI agents that handle real tasks. This isn’t a pitch deck — it’s a working system.

The architecture is built for trust. Our dual-agent architecture separates execution from governance by design. Agents do the work. Mission Control gives humans visibility and control. Every sensitive decision routes through oversight with full audit trails. In financial services, this isn’t optional — it’s the reason clients say yes.

You’ll build against real systems from day one. Through our sister companies in life insurance, payments processing, loan origination, and financial systems, you’ll have direct access to production data, live workflows, domain experts, and real compliance requirements. No waiting months for a design partner — the design partners are already in the building.

The domain is hard — and that’s the point. Regulated financial services, compliance-heavy industries, organizations where “move fast and break things” gets you a consent order. If AI can work here — with source verification, bias detection, and audit trails — it can work anywhere. Easy problems attract average engineers. Hard problems attract the best.

Atlanta is having a moment. The city’s tech scene is growing fast, cost of living is rational compared to the coasts, and Buckhead is one of the best neighborhoods in the Southeast. If you’re in Atlanta already, great. If you’d consider it, even better.

 

Peach Pilot is an equal opportunity employer. We evaluate candidates based on their abilities, experience, and potential — nothing else.