AI Harness Engineer_XC

Role Summary

• AI Agent enterprise scenario mining and implementation
• Go deep into the business department or customer site to understand the business flow of the enterprise in operations, R&D, O&M, testing and other scenarios.

process, data foundation and actual pain points.

• Identify high-value business scenarios suitable for AI Agent implementation, and evaluate business benefits, technical feasibility, data availability,

Implementation complexity and priority.

• Design AI Agent solutions for real enterprise processes, including but not limited to:
• Enterprise knowledge assistant
• Intelligent customer service Agent
• Data Analysis Agent
• Research and develop efficiency-improving Agents
• Office Automation Agent
• Approval flow/work order flow Agent

-Multi-Agent collaborative business assistant

• Responsible for promoting the entire process of AI Agent from PoC, MVP to production launch, ensuring that the solution can be used in real business environments

Medium and stable operation.

• Collaborate with business, production research, operation and maintenance, security and other teams to promote the integration of Agent applications into existing enterprise systems, such as IM,

Work order system, knowledge base, data platform, business system, etc.

• Establish an AI Agent implementation effect evaluation mechanism, focusing on accuracy, task completion rate, response delay, usage rate, and human efficiency.

Indicators such as savings, cost benefits, and business conversion are continuously optimized.

• Summarize the typical models, problem lists, implementation methodologies and reusable cases in the enterprise implementation process to form a standard

Deliver assets.

• AI Agent underlying capability building
• Participate in the construction of enterprise-level AI Agent technology base, including:
• Agent orchestration framework
• Prompt management

-Context management

• Memory mechanism
• Tool call
• Plug-in system
• Mission planning
• Multiple rounds of dialogue

-Multi-Agent collaboration

• Build and optimize RAG capabilities, including:
• Document parsing
• Document slicing
• Vectorization
• Recall strategy
• Reorder
• Knowledge update
• Permission isolation
• Traceability of answers
• Illusion control
• Design and implement the connection capabilities between Agent and enterprise tool systems, including:
• API calls
• Database query
• Workflow trigger
• Report generation
• Message notification
• Document handling
• Automated execution
• Build enterprise-level Agent workflow orchestration capabilities to support conditional judgment, manual confirmation, and exceptions in complex business processes

Digging into the bottom line, multi-role collaboration and long-term task execution.

• Responsible for the construction of the evaluation system for large models and Agent applications, including:
• Prompt review
• RAG Review
• Task execution evaluation
• Automated regression testing
• Analysis of failure cases
• User feedback closed loop
• Build Agent observability and operational capabilities, including:
• Log tracking
• Call chain analysis
• Cost statistics
• Model effect monitoring
• Abnormal alarm
• Run quality analysis
• Participate in enterprise-level AI security and governance capacity building, including data permissions, sensitive information protection, content security, and operations

Auditing, model call compliance, and security boundary design.

• Continue to track new technologies in large models, AI Agent, RAG, multi-modality, code generation, automated execution, etc.

And promote engineering verification and implementation in enterprise scenarios.

• Solution and delivery support
• Output AI Agent solution documents, technical architecture diagrams, PoC solutions, and implementation to customers or internal business teams

Plan and go-live acceptance criteria.

• Participate in delivery or internal digital projects and provide technical solution evaluation and implementation path design for complex business scenarios.
• Liaise with customers or business side technical teams to complete system integration, interface joint debugging, data access, permission configuration, and online launch

Deployment and troubleshooting.

• Addressing issues such as unstable effects, inaccurate knowledge, uncontrollable processes, and complex system integration during the implementation of Agent.

problems and provide engineering solutions.

• Precipitate common components, templates, Agent configuration specifications, Prompt templates, knowledge base construction specifications, test specifications and

Deploy manuals to improve the efficiency of subsequent project delivery.

• Provide training to business users, implementation teams, or customers to help them understand, use, and operate AI Agent capabilities.

Requirements

1. Technical capabilities
2. Bachelor degree or above, preferably in computer, software engineering, artificial intelligence, data science, automation and other related majors

First.

1. Familiar with the large model application development process and understand LLM, Prompt Engineering, Function Calling/Tool

Basic capabilities such as Calling, RAG, Embedding, and vector databases.

1. Be familiar with at least one AI Agent or large model application development framework/platform, such as:

• LangChain
• LlamaIndex
• LangGraph
• Dify
• Coze
• FastGPT
• AutoGen
• CrewAI

Semantic Kernel

• Have solid back-end development capabilities, be familiar with at least one language among Python/Java/Go/Node.js, and be able to work independently
• Complete API service, business logic, data processing and system integration development.
• Familiar with common databases and middleware, such as MySQL, PostgreSQL, Redis, Elasticsearch, Milvus,
• pgvector, Kafka, message queue, etc.
• Familiar with enterprise system integration methods, including REST API, Webhook, OAuth, SSO, database connection, file decryption
• Analysis, authority system, etc.
• Understand engineering capabilities such as model deployment, inference services, containerization, CI/CD, log monitoring, etc., and have experience in actual production environments
• Online experience is preferred.
• Candidates with experience in Agent evaluation, RAG effect optimization, Prompt optimization, and model calling cost optimization will be given priority.

Other information

1. Internal Referral bonus of this vacancy: RMB 3,000 (Valid only for Bosch associates). For the detailed regulation, please refer to Bosch China Internal Referral Policy

About the Company

Do you want beneficial technologies being shaped by your ideas? Whether in the areas of mobility solutions, consumer goods, industrial technology or energy and building technology - with us, you will have the chance to improve quality of life all across the globe. Welcome to Bosch.