InfoVision Inc. is seeking an AI Engineer to design and develop the VZ Application AI Modernization Factory, an AI-powered platform aimed at modernizing large-scale enterprise legacy applications. The role involves driving the technical vision of the AI Factory, collaborating with various teams to enhance the platform's capabilities and efficiency.
Responsibilities:
- Implement a prompt engineering system using structured YAML and Markdown templates, including: Dynamic placeholder substitution, Priority filtering, Category-based routing, Multi-instance LightRAG targeting
- Build and enhance the Adaptive Questioning Framework, featuring: LLM-driven recursive questioning, Configurable probing depth and levels, SQL indirection detection, Migration-critical validation guarantees
- Implement and maintain MCP server integrations, including: Vector store operations (upsert, search), Neo4j graph database queries, File metadata retrieval
- Design, build, and maintain a VS Code extension (TypeScript/Node.js), including: Chat participant integration, Command handlers, Guided conversational workflows
- Design and implement a multi-stage modernization pipeline: Application selection, Module-level targeted analysis, Adaptive deep-dive questioning, LLD (Low-Level Design) generation, Code instruction generation, Test instruction generation, Implementation guidance
- Develop and evolve a modular extension architecture, including: Services layer: LLM, session, file, user, adaptive questioning, Handlers: Chat participant, conversations, APIs, workflows, Utilities: Embeddings, token management, error tracking, SQL detection, UI components: Buttons, markdown rendering, progress indicators
- Implement a tiered error-handling framework: Early-stage failure: Stop execution and prompt connectivity diagnostics, Mid-stage failure: Pause and auto-retry with exponential backoff, Late-stage failure: Continue with partial results, Error classification: NETWORK, AUTH, SERVER, TIMEOUT, UNKNOWN
- Maintain build and packaging pipelines, including: TypeScript strict compilation, Bundling, Automated VSIX packaging
- Integrate the VS Code extension with LightRAG services, including: Connection lifecycle management, Endpoint targeting and routing, Contextual retrieval of legacy code artifacts
- Collaborate with: LightRAG platform teams on ingestion pipelines and retrieval quality, AI engineering peers on shared architecture and enhancements
- Maintain Python-based services for vector operations, including: Cosine similarity, Batch similarity computation, JSON-based TypeScript ↔ Python subprocess interoperability, Automatic TypeScript fallback on failures
- Manage embedding pipelines, including: External embedding API integrations, Batch processing, Exponential backoff retry strategies, Configurable batching