Back to Home
Senior Backend Engineer

Jia Wei Ho

Distributed Systems, Enterprise Search & AI

Stockholm, Sweden
Jia Wei Ho

Professional Summary

Senior Backend Engineer specializing in Kubernetes-deployed services, decoupled architectures via event-driven messaging (Kafka), hybrid search implementation (Elasticsearch/Vector/Lexical), and production-ready Agentic AI/LLM systems. Passionate about engineering high-performance, resilient, and observable large-scale Kotlin/Java ecosystems with automated evaluation pipelines.

Core Competencies

Backend & System Design

KotlinJavaDropwizardGraphQLMicroservicesEvent-Driven Architecture (Kafka)Redis

Agentic AI

Multi-Agent Orchestration (Handoff & Supervisor)LangChain4jOutput GuardrailsTool Argument GuardsPrompt Engineering (Handlebars/Hot-reloading)Mockable LLM Factories

Search & Analytics

Elasticsearch (Vector/kNN)Use Embeddings (ELSER/E5)Run LLM Evaluation (LLM-as-a-Judge)

Data & Observability

KafkaPostgreSQLOpenTelemetryElastic APMPrometheusTestcontainers

Professional Experience

Senior Backend Engineer

Volvo Support Platform

Sweden
2022 – Present

Key Achievements & Deliverables

Agentic AI & Multi-Agent Orchestration: Developed multi-agent solutions using Supervisor and Handoff patterns, incorporating template-driven system messages, custom tool setups, and LangChain for internal stakeholders and Volvo Support end-users.
Content Integrations: Developed and maintained continuous integrations, including Vehicle APIs, empowering content editors and stakeholders to produce high-quality support content with faster time-to-market.
Platform Migrations: Led multiple service migrations to align with enterprise engineering standards, modernizing infrastructure while reducing long-term operational costs.
LLM Guardrails & Security: Engineered anti-hallucination guardrails by integrating fuzzy matching against Elasticsearch retrieval results to validate generated responses.
Event-Driven Data Ingestion: Built a decoupled, event-driven Kafka pipeline for CMS synchronization (Contentstack, Volvo Content API, etc.) to reliably ingest and scale multi-source support content.
Enterprise Hybrid Search: Used Elasticsearch and Kibana query parameters to enhance search relevancy across vector and lexical retrieval methods.
LLM Evaluation Framework: Used internal evaluation systems utilizing deterministic path validation combined with LLM-as-a-Judge techniques and Redis memory analysis.
Testing & Observability: Implemented behavioral, outcome-based testing strategies to ensure systems are validated against functional output rather than execution paths alone.

Key Featured Projects

New Content Type Integration

Led cross-team architecture and API research to ingest a critical new content stream, driving a key corporate KPI to accelerate user query resolution.

Anti-Hallucination Guardrails

Spearheaded the development of a fuzzy-matching verification system for the Agentic AI chatbot, ensuring responses are anchored strictly in validated source articles.

Global Localization Pipeline

Improved language and locale handling across CMS content to deliver accurate, market-specific documentation globally.

Education

M.Sc. in Computer Science & Engineering

University • 2005 – 2011

Languages

Swedish

Native

English

Fluent

Let's Connect

Interested in discussing system design, backend engineering, or AI orchestration?