AI Agent Engineer

Lovefreedom Solution•Sunnyvale, CA

7d•Remote

About The Position

Sekai is a fast-growing consumer social AI startup building what can be described as the TikTok of interactive mini-apps. Sekai is creating a new way for users to turn natural language into playable, shareable digital experiences, powered by a proprietary coding agent. This is a rare opportunity to join one of the few U.S.-based consumer social startups scaling in the AI era. With 1M+ monthly active users, strong product-market signal, and $30M raised from a16z, Mayfield, and Khosla, Sekai is entering a major growth phase with a reported $150M post-money valuation and Series A momentum. As an AI Agent Engineer, you will own and set the technical bar for Sekai’s coding agent, the core engine that transforms user prompts into working interactive mini-apps. This is a high-ownership role for an engineer who has built real agentic systems in production and understands tool use, orchestration, repair loops, evaluation frameworks, and model strategy. Sekai is not building another AI wrapper. The coding agent is the product’s moat. You will be designing the system that turns natural language into playable experiences for more than 1M monthly active users. This means your work will directly shape how users create, interact, and publish AI-generated mini-apps at scale. The company is backed by top-tier investors including a16z, Mayfield, and Khosla, and is led by a CEO with deep consumer social experience as a former Director of Product Management at ByteDance. That combination of AI infrastructure, consumer product DNA, and real user traction makes this a rare seat for someone who wants to build at the edge of agentic systems and social creation.

Requirements

3–8 years of experience building and shipping production-level AI agentic systems
Experience building agentic systems involving tool use, orchestration, and repair loops
Experience shipping and operating complex systems in production
Strong proficiency in Python or TypeScript
Experience developing evaluation frameworks for AI or LLM systems
Familiarity with eval harnesses, regression testing, and failure taxonomies
Familiarity with modern agent frameworks such as LangChain, LangGraph, or similar tools
Strong systems thinking and the ability to design reliable, debuggable AI infrastructure
Bachelor’s degree in Computer Science or a related field
Native or native-level Mandarin proficiency
Ability to work remotely on Pacific Time hours
Availability for a Sunday evening standup

Nice To Haves

Experience at a VC-backed AI startup
Experience at a coding agent, developer tools, or IDE assistant company
Hands-on experience building a coding agent, devtool, or AI-powered IDE assistant
Experience with model routing, benchmarking, latency optimization, and reliability testing
Deep familiarity with agent frameworks, AI infrastructure, and the current open-source agent ecosystem
Track record of building AI systems used by real users in production

Responsibilities

Design and own the agent runtime and orchestration layer for Sekai’s coding agent
Build long-horizon agent workflows across prompt, planning, generation, execution, validation, repair, and publishing
Develop reliable agent loops involving tool use, orchestration, error handling, and repair flows
Build robust evaluation systems, including eval harnesses, regression suites, quality checks, and failure taxonomies
Design model strategies, including routing, benchmarking, reliability improvements, and cost/latency optimization
Create debuggable agent systems with tracing, metrics, alerts, and observability
Improve agent performance, reliability, and user-facing output quality
Work across product, engineering, and AI systems to turn natural language into polished interactive experiences
Stay current on modern agent frameworks, open-source tools, model releases, and evaluation tooling
Help set the technical bar for agentic development in a fast-moving consumer AI environment