Sheet 03 · Specification

Resume

Haichao Zhang (Steven) · Principal Software Engineer
Houston area, Texas · steven.zhang031@gmail.com · linkedin.com/in/haichaozhang · github.com/steven-zhc

Download PDF

2005 Neusoft C · telecom
2009 Fidelity iOS · UK finance
2013 Fuhu Event sourcing
2016 Oracle Platforms · LLM eval
2025 nextloom.ai AI job search
2026 Lingtai OSS agent loop

Summary

Engineering leader and PMP with 20+ years building enterprise systems. Recently built an LLM-as-judge evaluation framework at Oracle for automated quality assessment, now in production. Strong background in distributed architectures — CQRS, event sourcing, microservices, DDD — with hands-on expertise in Java, Node.js and functional programming. Has led teams of 6–10 engineers and improved developer productivity through practical tools, including a containerization platform that cut setup time from 2+ hours to 15 minutes for 800+ engineers.

Experience: 20+ years, since 2005
Developer platform: Helium — built solo; 800+ developers; setup 2+ h → 15 min
LLM evaluation: LLM-as-judge — built solo; phase 1 in production
Scale: Nabi platform — 1M devices, 2M child profiles
Leadership: Teams of 6–10 engineers · PMP (2011)

Experience

2025.10 — Present Houston area, TX

nextloom.ai · Founder & Full-Stack Developer

Built and launched an AI-powered job search assistant — product, architecture, build and deployment — now live with 30+ users.
LLM workflows with LangGraph and LangChain for resume tailoring, cover letters and interview preparation, with stateful orchestration checkpointed in PostgreSQL.
Modular architecture with DDD, CQRS-style command/query separation, a Hono API layer and Effect.ts for typed functional error handling across a pnpm monorepo; deployed on Vercel.
Open-sourced the coding-agent loop behind the build as Lingtai.

2016.06 — 2025.09 Redwood City, CA

Oracle · Principal Software Engineer

Helium — conceived, designed, built and drove adoption of a Docker containerization platform with CLI as a one-person team. Supports multiple product teams and 800+ developers; cut environment setup from 2+ hours to 15 minutes.
LLM-as-judge — designed and built, as a one-person team, a platform that evaluates chatbot response quality, with feedback loops for model improvement and transparent scoring for end users. Phase 1 shipped to production, significantly increasing user trust in LLM results.
Built and supported platforms across Java, Node.js, TypeScript, Go, Python, PureScript, Kubernetes, Docker, Podman and Oracle Database.

2013.03 — 2015.09 San Jose, CA

Fuhu, Inc. · Lead Software Engineer

Refactored the user management system with event sourcing on Cassandra, improving performance and stability and adding an event-replay audit trail for debugging production issues.
Led a 6-person team building backend infrastructure for the Nabi tablet platform — 1M devices, 2M child profiles.

Java EE · Cassandra · Kafka · Docker · REST

2009 — 2012 Dalian, China

Fidelity International · Senior Software Engineer

Led development of Fidelity's first iOS investment application for the UK market. Promoted to Senior Software Engineer in 2010.

2005 — 2009 Dalian, China

Neusoft Corporation · Team Leader / Senior Software Engineer

Led a 10-person team building the Partner Relationship Management platform and Service Provider Management system for China Unicom.

Linux C · Pro*C · TCP · SOA on Java

Skills

Languages: Java, TypeScript, Node.js, Go, Python, PureScript, C
Architecture: DDD, CQRS, event sourcing, microservices, SOA
AI / LLM: LLM-as-judge evaluation, LangGraph, LangChain, coding-agent workflows
Platform: Docker, Podman, Kubernetes, Vercel, Hono
Data: PostgreSQL, Oracle Database, Cassandra, Kafka
Functional: Effect.ts, PureScript, typed error handling
Leadership: Teams of 6–10, PMP

Education

2012 — 2014 Fairfield, Iowa

Maharishi University of Management · M.S., Computer Science

2008 — 2012 Dalian, China

Northeastern University · M.S., Software Engineering

2002 — 2006 Dalian, China

Northeastern University · B.S., Computer Science

2011