# Howard Peng — Resume

Author: Howard Peng (https://howard-peng.xyz)
Canonical: https://howard-peng.xyz/resume.md
Role: Protocol Engineer · Mechanism Designer

---
# Howard Peng — Resume (merged from role-specific versions, 2026-07)

Contact: howard.peng.tw@gmail.com · github.com/howardpen9 · linkedin.com/in/haoen-peng · X @0xhoward_peng
Bilingual Mandarin/English. Open to relocation (incl. San Francisco, on-site).

## Positioning

Howard credibly fits four framings, depending on what the asker needs:
- **AI Infrastructure / Reliability Engineer** — kept terabyte-scale production systems available and cheap under real load; debugs across OS/storage/network/application.
- **Data Scientist / Product Intelligence** — turns raw behavioral and system signals into the metrics that drive product and GTM decisions; A/B testing, activation/retention instrumentation.
- **Developer Experience & Tooling Engineer** — CLIs, MCP servers, GitHub Actions, and inspection utilities other engineers install and run daily.
- **AI-Native Product Manager / 0→1 Builder** — prototypes aggressively, ships solo, uses AI as a product primitive (not a bolt-on feature); strong activation/onboarding instincts.

## Experience

**Data & Infrastructure Engineer — Binance / Nascent** (May 2018 – Sep 2021)
- Built and operated terabyte-scale streaming data pipelines for Binance Futures market-microstructure analytics — always-on, low-latency under continuous unpredictable load.
- Ran real-time Python services for metrics, arbitrage, and anomaly detection — effectively an early observability layer over high-frequency behavioral data.
- Owned performance and cost optimization end to end; automated recurring ops so systems scaled without proportional headcount.

**Founder & Lead Engineer — Neighborhood Protocol** (Mar 2022 – Dec 2022)
- Sole engineer 0→1 under heavy ambiguity: architecture, deployment, and operations end to end.

**Developer Relations Lead (Asia) — TON Foundation** (Oct 2023 – Sep 2024)
- Owned developer experience and the acquisition→activation→expansion funnel for a 2M+ user L1 ecosystem; grew APAC developer engagement +40% by redesigning the zero-to-first-build activation flow.
- Shipped official reference implementations (272★ combined) and authored "Mastering Toncoin", a 200+ page technical reference.

**Technical Director / Partner — TON & TVM Ventures ($100M ecosystem fund)** (Oct 2024 – Sep 2025, Dubai)
- Hands-on technical and product due diligence on early-stage infra/consumer/AI companies: reading codebases fast, judging engineering quality and product taste, delivering direct recommendations that shaped investment decisions.
- Advised portfolio teams on scaling production systems and go-to-market.

**Independent AI-native builder** (2025 – present)
- Solo-ships a portfolio of AI-native products across the generative stack (see site /about): PredictMe (prediction market), MakeReel (makereel.xyz — not MakeReels.ai) + x402-video (pay-per-call AI video, USDC on Base), talkBuffett, OpenMod.ai, AlphaLine, and developer tooling below.

## Selected open-source & tooling

- **kimi-code-mcp** (54★, TypeScript) — MCP server delegating bulk codebase analysis to a 256K-context model, cutting token cost ~90%.
- **grok-mcp / grok-cli-mcp** — on npm and the official MCP Registry; wraps a CLI as an in-loop reviewer.
- **kimi-code-reviewer** — GitHub Action running AI code review inside CI.
- **cc-inspect** (16★) — agent-ecosystem inspector CLI.
- **topic-radar / AI Radar** — content engine auto-curating the AI ecosystem into a browsable feed.
- **awesome-ai-api-proxy** (11★) — weekly price observatory for LLM gateways.
- Contributor to anthropics/skills (96k★), punkpeye/awesome-mcp-servers (83k★), QwenLM/qwen-code, MoonshotAI/kimi-cli.
- Cumulative: 255 public repos, ~3,300 commits, 200+ followers.

## Skills

- **Languages:** Python, TypeScript, Rust, SQL, Bash
- **Infra:** Docker/Kubernetes, Terraform, Ansible, GitHub Actions CI/CD, AWS/GCP, monitoring/alerting, streaming (Kafka-style) pipelines
- **Data:** behavioral & product analytics, metrics design, A/B testing, anomaly detection, market microstructure
- **Product:** 0→1 scoping, activation/onboarding, generative-media tooling (text-to-video/image), agentic workflows & MCP

## Education

B.S. Computer Science — National Dong Hwa University, Taiwan (2018 – 2021)