电鸭
5人浏览/8人投稿
8小时前
待托管赏金
【About Us & The Opportunity】
We’re building the Context OS for AI agents — the foundation layer that makes large language models truly reliable at scale. LLMs today are powerful but brittle; their performance degrades as context expands. Our platform dynamically optimizes context in real time, improving both accuracy and latency for LLM features—automatically.
Backed by Andreessen Horowitz (a16z) and already serving over 10% of the Forbes AI 50, we are redefining AI infrastructure. We believe the future isn't bigger models, but smarter context.
To power this OS, we are hiring a Founding Staff Engineer in China. This is a hands-on technical powerhouse role for a pure Individual Contributor (IC) who will architect, code, and ship mission-critical features in TypeScript / Next.js from Day 1. You will work directly with the U.S. founders to build our core architecture, focusing on sub-second LLM response times and rock-solid reliability. You will earn leadership through technical mastery, eventually mentoring a small, elite China-based team.
【Core Responsibilities】
1. Hands-On Development (35%)
Personally write, review, and ship complex features in TypeScript / Next.js.
Debug, refactor, and optimize mission-critical pipelines for performance and scalability.
Push production code daily — your commits define velocity.
Solve deep technical challenges around latency, model streaming, and concurrency.
Collaborate directly with U.S. engineers to design elegant, maintainable codebases.
Scenario:
You discover the existing RAG implementation is bottlenecked at inference. You personally rewrite the TypeScript data fetcher and introduce streaming logic to cut response time by 40%.
2. Independent Technical Architecture (25%)
Design systems that meet global-scale demands for reliability, uptime, and low latency.
Architect caching, database, and edge strategies using Redis, Postgres/DynamoDB, Cloudflare, and AWS Lambda.
Make critical architectural calls without waiting for approval — you’re trusted with full technical judgment.
Ensure design patterns align with the company’s single-codebase model for global consistency.
Scenario:
When a U.S. founder flags scaling concerns, you propose a caching + edge rendering architecture that slashes cost per API call by 25% while improving latency for China users.
3. Product & LLM Application Expertise (20%)
Build and iterate on LLM-powered product features (prompt orchestration, retrieval, streaming).
Experiment with RAG, tool use, and function calling.
Use evaluation frameworks like Langfuse or Braintrust to measure prompt performance.
Balance model accuracy, latency, and cost with a data-driven mindset.
Scenario:
You run an A/B test between vanilla prompts and RAG-enhanced versions, achieving a 20% improvement in factual accuracy with zero latency penalty.
4. Team Leadership & Mentorship (10%)
Grow into leadership. You’ll start as the lead IC, then mentor and support 2-3 local engineers.
Conduct code reviews, provide architectural guidance, and elevate team standards.
Partner with U.S. leadership on hiring, onboarding, and performance feedback — without full managerial overhead.
Instill a culture of technical rigor, ownership, and curiosity.
Scenario:
A junior dev struggles with concurrency in Next.js. You pair program, teach them proper async patterns, and document the fix as a shared learning post for the team.
5. Cross-Border Communication (5%)
Communicate fluently with the U.S. team in English across Slack and Zoom.
Translate product requests into actionable technical tasks.
Provide daily asynchronous updates and weekly sprint summaries.
6. International / U.S. Startup Experience (5%)
Bonus if you’ve worked with U.S. startup teams and understand rapid iteration, ownership, and lean product cycles.
Comfortable navigating async work, time zone differences, and ambiguity.
【Qualifications】
Must-Have:
Expert-level fluency in TypeScript and Next.js; you are a hands-on builder who codes daily, not just a reviewer.
Proven experience architecting scalable, latency-sensitive systems using AWS, Redis, and serverless stacks.
Strong understanding of LLM application development (prompt engineering, streaming, RAG).
Fluent in English for daily cross-border communication.
Experience mentoring or leading small pods, or a strong desire to grow into a manager role.
Based in Mainland China (Beijing/Shanghai preferred)
Nice-to-Have:
Previous U.S. startup or early-stage experience.
Comfortable navigating async work, time zone differences, and ambiguity.
Exposure to Langfuse, Databricks, or ClickHouse.
Experience optimizing LLM API performance.
Prior work in latency-sensitive systems (e.g., Bytedance, Tencent).
【Compensation & Growth】
Base Range: $100K–$150K USD .
Structure: Independent Contractor (preferred) or EOR conversion.
Path: This will be the highest-paid engineer in China within the company, with a clear growth path to team leadership.
