Yulin Deng

Senior AI Platform Engineer | Agentic Systems · RAG · Python Backend

Shanghai, China | 17679329004 | 1016068291@qq.com | GitHub

Summary

AI platform engineer with 8+ years of experience building Python backend services and enterprise platforms, including nearly 2 years delivering production RAG, agentic, and MCP systems. Experienced across multi-source ingestion, multimodal document processing, retrieval and reranking, agent orchestration, long-term memory, authorization, asynchronous workloads, and observability. Core contributor to the 2.8k+ star Jarvis Registry open-source project. Reduced data-sync latency from hours to minutes, cut query latency by approximately 30%, and delivered an enterprise access-governance solution that saves more than RMB 2 million annually.

Technical Skills

  • AI / LLM: RAG, multimodal RAG, LangGraph, LlamaIndex, MCP / A2A, tool calling, embeddings, query rewriting, reranking, OpenAI, AWS Bedrock, Azure OpenAI
  • Backend: Python, FastAPI, Django / DRF, Flask, asyncio, Celery, REST APIs, task scheduling, service refactoring, performance optimization
  • Data & Infrastructure: PostgreSQL, MySQL, SQL Server, Redis, Weaviate, MongoDB, RabbitMQ / SQS, Docker, Kubernetes / KEDA, AWS
  • Security & Observability: OAuth 2.0 / OIDC, JWT, RBAC / ACL, OpenTelemetry, Prometheus, SkyWalking, structured logging, distributed tracing

Professional Experience

Global AI Startup | AI Application Developer / AI Platform Backend Engineer

Jan 2025 - Present

  • Led the migration of an enterprise RAG service from LangChain to LlamaIndex, reorganizing it into layered API, ingestion, parsing, embedding, retrieval, reranking, and scheduling modules while unifying REST and MCP query paths.
  • Designed connectors for AWS S3, Google Drive, SharePoint, and SQL Server with full, incremental, scheduled, and event-driven synchronization; used Celery / SQS batching, retries, import deduplication, and status tracking to improve processing efficiency by approximately 50% and reduce sync latency from hours to minutes.
  • Built a multimodal processing pipeline for PDF, Word, PowerPoint, Excel, email, and image content, integrating OCR, layout and table extraction, image processing, and normalized attachment metadata for retrieval and source citation.
  • Implemented Weaviate metadata filtering, query rewriting, reranking, and source-level authorization, with model integrations across OpenAI, AWS Bedrock, and Azure OpenAI; reduced query latency by approximately 30% and runtime memory usage by approximately 25%.
  • Built MCP server and tool-calling capabilities with OAuth and JWT validation, enabling agents to access enterprise documents, Google Workspace, and databases within user authorization boundaries, supported by actionable logs and metrics.

Stack: Python / FastAPI / LlamaIndex / Weaviate / Redis / Celery / SQS / OpenAI / AWS Bedrock / Azure OpenAI / MCP

Shanghai JIDU Automotive Co., Ltd. | R&D Productivity Platform / Enterprise Integration Engineer

Jan 2022 - Dec 2024

  • Integrated 10+ internal systems, including Jira, Jama, Confluence, Feishu, and feedback, bug, and vehicle-issue platforms; built REST endpoints, Behaviours, Listeners, and automated workflows with ScriptRunner / Groovy, increasing process automation by approximately 85% across PMO, engineering, and supplier collaboration.
  • Designed a Webhook-to-MQ architecture that reduced direct Jira API load and improved scalability and fault isolation for cross-system event processing.
  • Connected MDM, Active Directory, HR, and ticketing systems to automate supplier account lifecycles and security-group access; reduced onboarding and offboarding from days to hours and saved more than RMB 2 million in annual license costs.

Stack: Jira / Jama / Confluence / Groovy / ScriptRunner / MQ / Elasticsearch / AD / LDAP / Feishu

Shanghai Shurui Information Technology Co., Ltd. | Python Backend Developer / Engineering Standards Lead

Feb 2021 - Jan 2022

  • Standardized Django / Flask project scaffolding and deployment practices, introducing consistent layering, coding conventions, code review, and Git Flow to improve delivery consistency and maintainability.
  • Refactored 5,000+ lines of core recruitment-system code and resolved 20+ legacy defects; reduced average API response time from 2.5 seconds to under 0.8 seconds through application and query optimization.

Stack: Python / Django / Flask / MySQL / Redis / Git Flow

Shanghai Data Exchange Co., Ltd. | Python Backend Developer

Feb 2020 - Jan 2021

  • Developed core business services, APIs, and system integrations; used asynchronous processing, modular refactoring, index optimization, and SQL tuning to improve throughput and reduce selected query times by approximately 50%.

Shanghai Yilu Investment Management Co., Ltd. | Backend / Systems Developer

Feb 2018 - Jan 2020

  • Built a hotel SaaS PMS and business platform from the ground up, including JWT-, Redis-, and MongoDB-based authentication and authorization services; implemented asynchronous payment polling and order-timeout handling with Celery and RabbitMQ for a product deployed at 200+ hotels.

Selected Projects

mate-assist | Automotive Sales and Operations Agent Backend

Private business project, 2026 | FastAPI / LangGraph / MySQL / Redis / Celery / Kubernetes / KEDA

  • Designed an 11-node LangGraph that handles context control, multimodal perception, query rewriting, intent routing, retrieval and grading, corrective re-retrieval, sales-script generation, and long-term memory extraction across knowledge, sales-support, discourse, and vision branches.
  • Implemented SSE streaming, time-to-first-byte and end-to-end latency metrics, node-level tracing, transient-failure retries, user cancellation, and MySQL-backed checkpoints and long-term memory, including conversation replay and answer regeneration.
  • Built Celery workflows for customer profiling, sales-funnel analysis, store briefings, lost-lead classification, and test-drive analysis, using exponential backoff, concurrency gates, cache short-circuiting, and KEDA autoscaling from zero to ten workers based on Redis queue depth.
  • Decoupled 50+ runtime prompt assets from the application image, enforced manifest, variable, and required-section validation as a CI gate, and used git-sync, immutable release tags, and startup checks for environment isolation and rollback.

Jarvis Registry | Enterprise MCP and A2A Agent Gateway

Core contributor | 2.8k+ GitHub stars | Nearly 300 authored commits on the main branch

  • Contributed across the enterprise MCP / A2A gateway, including server and agent registration, semantic discovery, reverse proxying, federation, the Skill Registry, and integrations for Cursor, Claude Desktop, GitHub Copilot, and VS Code.
  • Built multi-provider OAuth 2.0 / OIDC capabilities spanning Keycloak, Amazon Cognito, Microsoft Entra ID, and Google, with JWT and session validation, scope-based RBAC, object-level ACLs, and tool-level access control.
  • Contributed to MCP / A2A workflow orchestration, scheduled execution, node-level failure attribution and recovery, plus structured logging, OpenTelemetry tracing, Prometheus metrics, Tempo / Grafana integration, and controls for sensitive trace data.

Education

Jiangxi University of Science and Technology | B.Eng. in Electronic Information Engineering | 2013 - 2017

Updated September 2026