AI Full Stack System Architect (Remote)

About Gro AI: Re-defining "sales" with AI

Gro AI byFormer Huawei Cloud Asia-Pacific CDO Create. Build the next generationAI Sales Infrastructure

• Company website: https://thegro.ai/

• Product video:https://www.youtube.com/watch?v=wHwa2Q4QvAc

• https://www.linkedin.com/in/leo-nuo-jiang/

Mission: Build the best way to sell with AI.

Our customers: Covering top technology startups and listed companies in Asia and Europe and the United States.

This is an early core engineering position at Gro AI. You will join as a founding team member, reporting directly to the founder, and deeply involved in product direction, system architecture and engineering standards.

We are not looking for an engineer to execute requirements, but rather someone who can proactively define problems, make engineering judgments, and continuously deliver product results in a highly uncertain early environment.

You need to prove:

Three. What will you be responsible for

1. Build the core Agent system:

2. Design reliable workflow orchestration:Construction task planning, tool invocation, state management, HITL, retry, rollback, idempotence, and exception recovery mechanism.

3. Deliver end-to-end products:Based on GCP + Next.js to build a highly reliable asynchronous full-stack architecture, supporting thousands of concurrent users.

4. Building data and retrieval infrastructure:Handling large-scale LinkedIn data and corporate knowledge base, building Evidence-Grounded Retrieval and quality assurance system.

5. Establish evaluation and observability systems:Design product-specific evals、workflow monitoring、latency / cost tracking, and feed the results back into system optimization.

6. Boost AI-native engineering efficiency:Transform the warehouse into an AI Coding Agent that can collaborate end-to-end in the engineering field.

Four. We hope you have

Production-level Agent system capabilities

• Agent architecture and orchestration:Able to abstract business processes into State Machine and Agent Loop, handling loops, branches, concurrency, interrupt recovery, HITL, rollback and idempotence; familiar with mainstream frameworks such as LangGraph, CrewAI, AutoGen, Mastra or self-developed runtime, and able to clearly distinguish between demo and production systems.

• Able to abstract system capabilities into combinable and single-responsibility Tools / Skills; follow open standards such as Agent Skills (SKILL.md) and Model Context Protocol (MCP); and design clear I/O schema, error codes, and side effects boundaries.

• Context Engineering：Consider the Context Window as a scarce budget, design Retrieval, Memory, compression and context assembly links; implement at least one long-term memory or retrieval solution in a production environment, and be able to explain its failure mode and rollback strategy; have PII boundary and prompt injection protection awareness.

• Evals and self-execution:

• Programming language:Proficient in TypeScript and familiar with Python; not limited by language in agent framework, evaluation tool and rapid prototype scenario.

• Cloud and asynchronous architecture:

• Full-stack abilities:Familiar with Next.js / React and backend API design (REST / RPC, authentication, caching boundary), focusing on TTFB, observability, and system maintainability.

• High concurrency and reliability:Have thousands of user-level concurrent production experience; familiar with Redis, message queue, connection pool tuning; understand exponentiation, retry, DLQ, rate limiting and circuit breaking; be able to make engineering trade-offs between delay budget, token cost and inference latency.

• Harness Engineering：Can transform the warehouse into an agent-friendly work space, including refined AGENTS.md, structured docs/, custom linter, structural testing and observability hooks.

• AI-augmented Development speed:Daily familiar with Claude Code, Codex, Cursor, etc. AI-assisted programming tools, able to review, debug and iterate agent outputs.

Working style and comprehensive quality

• Ownership & High Agency：Able to independently decompose fuzzy problems into online product solutions after the founder gives direction, actively defining scope, priority and delivery rhythm.

• Bias to Ship：Able to make reasonable judgments between speed and quality, quickly publish, and continuously iterate through real user feedback.

• Customer Empathy：Willing to directly communicate with early customers, participate in forward-deployed prototype, problem localization and fault diagnosis.

• Low-ego Collaboration：Able to collaborate intensively with Founder, researchers, and designers in a small team, openly accept feedback, and quickly adjust when the direction changes.

• Product thinking:Don't blindly pile up code, but think about technical implementation from the perspective of product value, user experience, and long-term retention.

• Globalization communication:Able to read complex technical documents, attend meetings with global customers, and write public technical content in fluent English; able to collaborate efficiently with teams in Asia, Europe, and North America.

• Research-aware:Continuously follow up on Anthropic, OpenAI, DeepMind, etc. the top laboratory's engineering blog and representative papers; stay sensitive to new models and frameworks, and quickly prototype to evaluate their value in the production stack.

• Handling large-scale LinkedIn data, web scraping, anti-crawling and anti-pressure, entity resolution or record linkage, and stabilizing the related system in a production environment.

• We have delivered products in compliance with SOC 2, GDPR, PII processing or prompt injection protection, etc. security and safety contexts.

• In Anthropic Claude, OpenAI, MCP, Agent Skills, etc. ecosystems have made public technical contributions, such as GitHub, Hugging Face, personal blogs, technical speeches or open source agent projects.

• Familiar with voice agent, multimodal agent, computer-use, etc. emerging agent forms.

Six. Why join us

• Participate in defining the next generation AI Sales Infrastructure:Join us in using AI to reshape the way sales teams work.

• Directly work with the founder:

• Real scene verification:The platform has served customers in Asia, Europe, and North America, and your code will directly affect the daily work of thousands of salespeople, with a very short feedback loop.

• Deep impact on product and technology direction:As an early engineering member, you will participate in key architecture decisions, engineering culture building, and product roadmap design.

• Obtain competitive returns:We provide competitive salary and early stock options, and we expect to complete our seed round of funding in Q3.

Seven. Application method

• Resume

• GitHub Homepage (optional)

One sentence: We are looking for an engineer who builds with AI every day, rather than just writing business code like a programmer.