08/13/2026 updated

**** ******** ****
100 % available

Senior Full-Stack Engineer | React, TypeScript, Node.js, Python, Web3

Ningbo, China
Only remote
Bachelor of Engineering, Network Engineering - Guangdong Polytechnic College (2020-2024)
Ningbo, China
Only remote
Bachelor of Engineering, Network Engineering - Guangdong Polytechnic College (2020-2024)

Profile attachments

Toni_Senior_Backend_Engineer.pdf

About me

Full-stack engineer with 5+ years of experience building SaaS, Web3, data and AI-enabled products. I deliver React/Next.js interfaces, Node.js/NestJS and Python services, PostgreSQL systems, APIs, cloud deployments and reliable 0-to-1 platforms.

Backend Engineering
Expertise in TypeScript/Node.js, NestJS, Python, Flask, SQL, PostgreSQL, MySQL, Prisma ORM, REST APIs, OpenAPI/Swagger, relational schema design and data contracts. Delivery of full-stack protocol platforms including smart-contract integration, wallet-based transaction flows and verifiable AI-asset exchange systems.

Distributed Systems & Transaction Infrastructure
Deep knowledge of transaction state machines, idempotency requirements, replay-safe transitions, retry and fallback design, RPC failure handling, reconciliation, failure taxonomy and operational metrics. Design of deterministic transaction lifecycles covering trade intent, escrow, oracle evidence, release, refund, dispute escalation and manual override.

Trading & Digital Assets
Working knowledge of market microstructure, order types and routing concepts, VWAP, rolling backtests, wallet authentication, EVM and Solana account concepts, smart-contract escrow, settlement and oracle validation. Implementation of Python research pipelines using VWAP-based features and rolling-window backtesting for crypto intraday-return signals.

Performance & Testing
Application of mixed precision, adaptive batching, memory monitoring, modular checkpoints, path autodetection, reproducible train/inference pipelines, technical acceptance criteria and failure-path validation to resolve constrained-GPU memory failures and ensure system reliability under resource constraints.

Data & ML Systems
Experience with PyTorch, FATE federated learning, Transformer recommenders, RQ-VAE, multimodal fusion, Pandas, NumPy, scikit-learn, EMD-ARIMA/SVD and model evaluation. Delivery of semantic-ID recommendation pipelines combining visual, text and collaborative signals.

Industrial & Integration Systems
Research and delivery experience in digital twins, predictive maintenance, multi-robot coordination, IoT/Edge architecture, platform-specific adapters and cross-system data workflows. Co-development of invention-patent filings in intelligent manufacturing and fault-prevention scheduling.

Cloud & Delivery
Proficiency in Docker, GitHub Actions CI/CD, Git/GitHub, AWS EC2, S3, RDS and Lambda, Vercel, release criteria and implementation sequencing for production-grade deployments.

Security & Leadership
Knowledge of RBAC, local-first privacy, differential privacy, homomorphic encryption, Sybil and referral-fraud controls, 0-to-1 product ownership, architecture documentation and cross-functional delivery.

API & Integration Architecture
Authoring of OpenAPI contracts, relational schemas, sequence flows and state machines. Building of platform-specific adapters and resilient handling for late, incomplete or conflicting upstream data. Authored 15+ OpenAPI endpoint specifications and 13+ user stories with technical acceptance criteria.

Experience

Professional experience5 - 10 years
Freelancer since2020

Languages

ChineseNative speakerEnglishNative speaker

Project history

Product Manager & Full Stack Engineer - Exchange, Escrow & Verification Infrastructure

Tokenta
Working on a verifiable AI-asset exchange and trust-minimised transaction platform. Designed the service and data architecture for wallet identities, asset records, proof submissions, transaction states, disputes and reputation. Built across the full-stack protocol platform with TypeScript, NestJS, PostgreSQL, Prisma, Docker, smart-contract integration, Wagmi, Ethers.js and RainbowKit. Defined a deterministic transaction lifecycle covering trade intent, escrow states, oracle evidence, release and refund paths, failure handling, dispute escalation and manual override. Structured the verification pipeline for automated endpoint testing, quota validation and proof-hash records. Established delivery and operating metrics for transaction completion, validation accuracy, latency, escalation, model regression and failure-reason analysis.

Senior Product Manager - Backend Platform & On-Chain Integration

ALLWEB3
Three-portal Web3 platform with oracle-verified KPI settlement. Architected a three-portal backend model for creators, projects and administrators, integrating smart-contract escrow, oracle-verified KPIs, RBAC authentication and transaction-state management. Authored 15+ OpenAPI endpoint specifications and 13+ user stories with technical acceptance criteria, ER diagrams, sequence flows and state machines; improved sprint velocity by 35%. Designed replay-safe settlement transitions with evidence requirements, dispute escalation and deterministic fund-release conditions tied to on-chain attestation. Worked with NestJS, PostgreSQL, Prisma, Swagger/OpenAPI, Docker, GitHub Actions and AWS infrastructure.

Product Designer & ML Engineer

Tencent Advertising Algorithm Competition
Multimodal generative recommendation and reproducible ML infrastructure. Built a semantic-ID recommendation pipeline combining 4,096D visual, 1,024D text and collaborative signals with RQ-VAE, cross-modal attention and a causal Transformer. Improved system reliability under resource constraints through mixed precision, adaptive batching, memory monitoring, modular checkpoints, path autodetection and consistent training/inference preprocessing.

Product Designer, Full-Stack Engineer & Product Manager - Robotics, Data & Web3

CyberOrigin
Teleoperation, data-platform and on-chain product systems. Owned architecture, UX and full-stack delivery for a responsive Next.js/React/TypeScript platform with Three.js hand-data visualisations, modular product services and technical case studies. Specified a modular teleoperation platform spanning IMU, optical and EMG capture, robot execution, RGB-D perception, low-latency control, device administration, multimodal data review and third-party APIs. Owned transaction and risk workflows across wallet onboarding, smart-contract settlement, tokenomics, analytics instrumentation and two rapid product iterations. Built security and abuse-control requirements covering Sybil resistance, referral fraud, address clustering, incentive abuse, escalation rules and user-facing recovery states.

Solo Product Designer & Full-Stack Engineer

JobBot Pro x Wenxiaoyan
Privacy-first job-application web app and Manifest V3 extension. Built dynamic form detection and one-click autofill across React, Vue and Angular sites, with Workday, Greenhouse and Lever adapters, MutationObserver-driven handling, fuzzy matching and local-first privacy controls.

Team Lead

Web3 Quantitative Trading Model
Crypto intraday-return prediction and rolling evaluation. Built a Python research pipeline using VWAP-based features, rolling-window backtesting and weighted Spearman rank correlation to evaluate intraday crypto-return signals.

Research Assistant - Systems Engineering & Digital Twins

Southern University of Science and Technology
Industrial systems, predictive maintenance and multi-robot coordination. Researched industrial digital twins and model-based systems engineering for automotive press and hot-stamping environments, including fault prevention, health-index formulation and dynamic scheduling. Co-developed two invention-patent filings including CN202310235924.9, and contributed to force/temperature-field modelling, steel microstructure processing and prognostics pipelines. Supported technical teams and student delivery including a national top-nine Industrial Internet and Digital Twin algorithm team.

Product & Full-Stack Lead - Federated Learning Risk Systems

WeBank / Shenzhen International FinTech Competition
DeepVise privacy-preserving predictive-maintenance and financial-risk platform. Designed cross-organisation data and model workflows using FATE federated learning, differential privacy, homomorphic encryption, multi-source signals, dynamic aggregation and maintenance feedback loops. Connected health indicators to operational decisions through weighted envelope features, SVD and EMD-ARIMA prognostics, remaining-useful-life analysis, API contracts and a Flask monitoring prototype. Documented experimental results of up to 80% fault-prediction accuracy and a 14% maintenance-cost reduction; placed Global Top 10.

Certificates

Blockchain and Money (15.S12)

MIT Sloan School of Management

2025

Two Invention-Patent Filings

SUSTech

2025

Bronze Award - Global Top 10

Sony Global University Innovation Challenge

2024

Global Top 10 - FinTech Competition

WeBank

2024

BEC Higher and IELTS Speaking 7.5

Cambridge English / British Council

2024

National Gold Award - Star Plan

AVIC International / Tianma

2023


Contact form

Log in to get in touch

You need to be logged in to use the contact form.

Sign upLog in