Software Engineer

Akia builds AI agents for hospitality — a system of actions that handles guest communication, drives direct bookings, manages upsells, and operates the full guest lifecycle.

Built on a modular skills architecture. The AI reasons, acts, and learns across every domain of hotel operations. The stack is real, the problems are hard, and the AI isn't a feature bolted on — it's the product.

We're still early, and there's a lot left to build.

We're building AI that can learn complex real-world operations and run them autonomously. That means designing agentic systems that compose reliably across domains, scale across wildly different environments, and make the right decisions with imperfect information.

This role is about building those systems. You'll design the architecture that lets autonomous AI act correctly in production at scale.

We believe the best engineering in the age of AI comes down to architectural judgment — knowing what to build, why the pieces fit together, and where the system will break under pressure. We use AI heavily in how we build, and we need engineers who can guide it well.

Design and build the skills architecture. Define how modular AI capabilities are structured, how they share context, and how they compose. Tool use, function calling, multi-step agent orchestration — production-grade and built to last.

Build the action and execution layer. The AI doesn't just converse — it acts. Action execution, rollback, retry, audit trails. Structured outputs, chain-of-thought reasoning, human-in-the-loop gating, and guardrails that control autonomy.

Architect the reasoning and retrieval pipeline. RAG pipelines, embedding strategies, context window management, semantic search. Build the systems that make agentic reasoning reliable and consistent at scale.

Design the AI harness and feedback systems. The software systems, environments, and feedback loops that govern how AI agents learn, improve, and operate safely. Evaluation frameworks, testing infrastructure, confidence calibration, and the engineering scaffolding that turns a language model into a reliable autonomous system.

Own scale and reliability. Production systems that handle real traffic across wildly different environments. Multi-tenancy, database performance, caching, horizontal scaling, observability, and the operational rigor that keeps autonomous AI safe in production.

CS degree or equivalent depth. Data structures, algorithms, distributed systems, networking, concurrency — as tools you reach for daily when designing real systems.

5–8 years building production systems. You've designed, shipped, and operated systems at scale. You've tuned database queries, debugged performance bottlenecks, managed migrations, and been on-call for systems you built. You know what breaks at scale because you've lived it.

Deep system design experience. End-to-end system architecture — you've designed entire systems, owned the tradeoffs, and evolved them over time. Distributed systems, database internals, caching, message queues, service decomposition, horizontal scaling, and reasoning about failure at every layer.

Genuine passion for AI agents. You're deeply curious about how autonomous systems work, how they fail, and what makes them reliable. You've built with LLMs — RAG, agent orchestration, tool use, prompt engineering, any or all.

Ability to guide AI, beyond just using it. You spot leaky abstractions, naive error handling, and unscalable data models in AI-generated code instantly. You know when to accept, redirect, or throw it away.

Critical thinking about AI output. You hold AI-generated code to the same bar as a senior engineer: correctness, performance, security, and behavior at the edges.

• Salary: $165,000 – $230,000
• Equity / stock options
• Full health, dental, and vision coverage
• Commuter benefits
• 401(k)
• PTO