Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Senior AI / ML and backend engineer with 13+ years of experience designing, building, and operating enterprise software, data platforms, automation systems, and production-grade GenAI applications. Strong hands-on background in LLM integration, RAG, semantic search, AI evaluation, prompt engineering, document intelligence, data pipelines, API/microservice architecture, cloud DevOps, and secure full-stack delivery. Experienced translating ambiguous business workflows into measurable AI automation, leading prototypes through production rollout, and communicating architecture, risk, adoption, and value to technical and non-technical stakeholders.…Senior AI / ML and backend engineer with 13+ years of experience designing, building, and operating enterprise software, data platforms, automation systems, and production-grade GenAI applications. Strong hands-on background in LLM integration, RAG, semantic search, AI evaluation, prompt engineering, document intelligence, data pipelines, API/microservice architecture, cloud DevOps, and secure full-stack delivery. Experienced translating ambiguous business workflows into measurable AI automation, leading prototypes through production rollout, and communicating architecture, risk, adoption, and value to technical and non-technical stakeholders.

WEN YUAN

AI Engineer, Back-End Developer, Software Engineer, +2





Senior AI / ML and backend engineer with 13+ years of experience designing, building, and operating enterprise software, data platforms, automation systems, and production-grade GenAI applications. Strong hands-on background in LLM integration, RAG, semantic search, AI evaluation, prompt engineering, document intelligence, data pipelines, API/microservice architecture, cloud DevOps, and secure full-stack delivery. Experienced translating ambiguous business workflows into measurable AI automation, leading prototypes through production rollout, and communicating architecture, risk, adoption, and value to technical and non-technical stakeholders.…Senior AI / ML and backend engineer with 13+ years of experience designing, building, and operating enterprise software, data platforms, automation systems, and production-grade GenAI applications. Strong hands-on background in LLM integration, RAG, semantic search, AI evaluation, prompt engineering, document intelligence, data pipelines, API/microservice architecture, cloud DevOps, and secure full-stack delivery. Experienced translating ambiguous business workflows into measurable AI automation, leading prototypes through production rollout, and communicating architecture, risk, adoption, and value to technical and non-technical stakeholders.

Available to hire

Senior AI / ML and backend engineer with 13+ years of experience designing, building, and operating enterprise software, data platforms, automation systems, and production-grade GenAI applications. Strong hands-on background in LLM integration, RAG, semantic search, AI evaluation, prompt engineering, document intelligence, data pipelines, API/microservice architecture, cloud DevOps, and secure full-stack delivery. Experienced translating ambiguous business workflows into measurable AI automation, leading prototypes through production rollout, and communicating architecture, risk, adoption, and value to technical and non-technical stakeholders.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Language

Work Experience

AI Software Evaluation Engineer at Meta

January 1, 2026 - Present

Evaluates production-scale AI-generated software solutions across Python and TypeScript repositories, focusing on correctness, maintainability, architectural consistency, testability, security, and production readiness. Designs multi-step agentic evaluation workflows for debugging, refactoring, repository reasoning, code synthesis, tool use, and autonomous software-modification tasks. Analyzes LLM failure modes such as hallucinated APIs, weak repository-context usage, incomplete implementations, brittle reasoning, unsafe refactors, missing tests, and incorrect integration assumptions. Creates structured rubrics for reasoning quality, instruction adherence, edge-case handling, code style, system-design alignment, and deployment feasibility, partnering with AI safety/model-quality teams to improve robustness and reliability.

Senior AI Engineer at NetEase

January 1, 2025 - December 1, 2025

Designed and delivered an enterprise RAG and semantic analytics platform using FastAPI, Airflow, Docling, OpenSearch, and LlamaIndex with embedding-based retrieval across 50K+ document chunks and structured records. Implemented hybrid retrieval (BM25 + vector similarity) with metadata filtering and source-aware generation, improving retrieval precision@5 by 35%+ versus keyword-only baselines. Built automated ingestion for parsing, chunking, embedding generation, index refresh, retries, and failure recovery, reducing manual refresh effort by 70%. Optimized chunking strategy and OpenSearch query patterns to keep common top-k retrieval under 500ms and reduce irrelevant context injection by 30%. Implemented LLM evaluation workflows for faithfulness, relevance, source grounding, citation coverage, hallucination detection, and fallback behavior, raising grounded-answer pass rate from ~70% to 85%. Added retrieval observability (traces, retrieved chunks, flags, failed-retrieval logs) to reduce

Full-Stack & AI Automation Engineer at ITLabPro

January 1, 2024 - October 1, 2024

Built an AI-enabled full-stack platform using TypeScript/Node.js/Express, MongoDB, JWT, and payment providers (Stripe/PayPal) supporting 1K+ user/project records and subscription workflows. Integrated LLM-powered semantic search, summaries, prompt-based assistance, support-style Q&A, and workflow acceleration, reducing manual lookup and repetitive support tasks by 40%. Developed asynchronous AI execution services with status polling, retries, result persistence, and failure recovery, improving workflow completion reliability to 95%+. Designed schemas for content metadata, user projects, generated outputs, prompt versions, payment events, and retrieval-ready documents to enable future RAG indexing. Implemented observability (prompt version tracking, IO logging, regeneration counts, quality flags, user feedback) and secure REST APIs with RBAC, validation, idempotency, and error handling across AI and payment workflows, reducing failed edge cases by 30–35%.

AI Evaluation & LLM Quality Engineering at Snorkel, Invisible, Alignerr, Outlier, AfterQuery

January 1, 2023 - January 1, 2026

Delivered LLM evaluation work across model response quality, coding-agent reliability, RLHF-style feedback, benchmark review, and software-engineering task evaluation. Evaluated AI-generated outputs for Python/TypeScript/JavaScript/React/Node.js, REST APIs, SQL, algorithms, debugging, testing, and system design, assessing correctness and production suitability. Built evaluation datasets, rubrics, prompts, and reviewer feedback processes for code generation, bug fixing, refactoring, repository reasoning, and multi-turn engineering workflows. Performed model ranking and hallucination detection with error taxonomy creation for measurable quality improvements across autonomous and tool-using agent workflows.

AI & Data Engineer at Creamistry

February 1, 2019 - July 1, 2022

Owned analytics automation for Creamistry China franchise operations, consolidating POS, delivery/orders, inventory, product sales, customer engagement, and campaign data into repeatable reporting workflows. Built Python/SQL ETL pipelines reducing manual Excel reporting by 60–70%, standardizing store/product/campaign metrics and improving weekly visibility into sales, demand, and channel performance. Designed KPI dashboards (revenue, order volume, average ticket, product mix, delivery vs in-store mix, refunds/cancellations, inventory risk, campaign ROI), reducing ad hoc reporting requests by 40%. Implemented data-quality rules for duplicates, refunded transactions, inconsistent product names, missing store identifiers, delayed exports, and date/timezone boundary issues, reducing discrepancies by 30%. Built forecasting and trend-analysis workflows for seasonal demand and inventory planning to identify stockout and underperforming items earlier. Supported COVID-period operating decisio

Software Engineer at Huawei

July 1, 2015 - December 1, 2018

Contributed to enterprise software systems implementing Java/Kotlin modules, Python/Node.js backend integrations, REST API workflows, authentication logic, database-connected services, and internal tooling. Delivered production-facing features including request validation, service orchestration, access control, session management, local persistence, API communication, and workflow automation across mobile and backend-connected systems. Integrated distributed services via REST APIs, improving reliability with retry behavior, pagination, error handling, and data consistency across environments. Managed SQL-backed data flows and debugging for authentication, sync, reporting, and integration scenarios, resolving partial updates, malformed payloads, duplicate submissions, and inconsistent records. Supported DevOps readiness with CI/CD validation, Linux troubleshooting, deployment verification, rollback planning, log analysis, and regression testing. Improved stability by handling network lo

Backend Engineer at ZBJ Network

February 1, 2013 - May 1, 2015

Built backend-focused marketplace modules with PHP/Laravel, JavaScript, SQL, and REST APIs for service listings, user profiles, order/workflow tracking, admin tooling, validation logic, and reporting views. Implemented deterministic business-rule workflows for order status changes, input validation, duplicate-submission prevention, user/session handling, and admin review processes. Improved marketplace reliability by resolving production issues including inconsistent order states, malformed inputs, SQL query errors, session edge cases, and broken admin workflows. Optimized frequently used backend routes and SQL queries, improving response times for common admin/marketplace workflows by 20%. Created reusable backend patterns for validation, data access, error handling, and workflow state transitions; collaborated on feature delivery, defect triage, release support, and production maintenance in a high-volume environment.