Job description
Translated by FDE Times from the original posting in Chinese.
About the team: Data Platform is ByteDance's data middle platform department. It provides one-stop big data solutions for the company's many business lines (including Douyin, e-commerce, livestreaming and local lifestyle services), covering the full pipeline of data production, cleansing, transmission, modeling and analysis, and offering solutions for diverse scenarios such as data development, experiment evaluation, user profiling and tagging, and augmented analytics. The Data Platform department is also committed to commercializing the data middle platform solutions that ByteDance has built up over the years, so that more industries can use our product capabilities to build their own data middle platforms. On Volcano Engine, we offer related product solutions such as the Marketing Growth Suite and the Data Middle Platform, and provide industry solutions for the broader internet sector, finance, automotive, new retail and other industries.
1. Technical lead for Agent projects: As the person with primary technical responsibility for Agent projects, you lead the full lifecycle from requirements research, architecture design, prompt engineering and Skills writing to launch and operations, and you are accountable for the customer's business outcome metrics; 2. Agent architecture design: Based on Volcano Engine's three foundation products (Interactive Assistant / Knowledge Engine / Analytics Agent), design end-to-end Agent architectures for customer scenarios, including multi-Agent orchestration, tool calling, knowledge chunking, RAG pipelines and Eval frameworks; be able to make architectural trade-offs and design alternative solutions when the foundation's capabilities fall short; 3. Deep involvement in the customer's business: Work with the customer's business teams to map out business processes, design where the Agent steps in, and identify the boundary between "where AI can step in and where humans act as the fallback"; understand the customer's business processes deeply enough to support technical decisions independently; 4. On-site technical decisions: In real-world situations where the product keeps iterating and customer needs keep changing, make technical architecture decisions, troubleshoot problems and adjust solutions independently, without relying on layer-by-layer approval from headquarters; 5. Collaboration across customers and roles: Identify scenarios common to multiple customers and turn them into reusable feature designs, code, Skills, prompt templates, Agent templates and Eval sets, leaving behind at least 1 cross-customer reusable asset per project; acting as the product R&D team's "on-site representative", give the product R&D team structured feedback on gaps in the foundation's capabilities found at customer sites, and take part in discussions on foundation product iteration priorities; along the Agent delivery chain, work with PM/BA/ETL/product R&D/CSM in a T-shaped collaboration, with clear ownership of technical decision-making and clear handoff standards.
1. Bachelor's degree or above in computer science, engineering, mathematics or a related field, or equivalent practical experience; 2. 5+ years of customer-facing technical architecture and delivery experience; proficient in Vibe coding; able to independently complete prototype development, customer communication and production-grade delivery; 3. Software engineering skills: proficient in Python (able to write production-grade code, not just analysis scripts), familiar with API development and basic frontend (able to build demos independently), and understanding of containerized deployment and CI/CD processes; 4. Data engineering skills: expert in SQL, with hands-on experience in data governance, data modeling and ETL, and an understanding of how critical data quality is to the performance of AI applications; 5. AI Ready data design skills: able to design a data layer that "can be consumed by Agents", with semantically clear metadata, context-friendly data structures and knowledge organization suited to RAG; experience with semantic layer/Ontology modeling is a plus.