Philip Cheung

AI Product & Transformation Lead

Financial Services | Fintech | Applied GenAI

MSc Computer Science, AI & Data Science (Merit)

London, United Kingdomgithub.com/pakmingcwww.Philip.pmLinkedInBook a 30-min chat

Professional Summary

AI Product and Transformation Lead with more than 10 years across financial services, fintech, quantitative markets and digital platforms, currently Head of AI at a Swiss corporate finance advisory firm. Translates commercial and financial-services problems into prioritised AI use cases, then owns product vision, MVP scope, roadmaps, requirements, evaluation criteria and governance through to adoption. The current portfolio spans an ultra-low-latency AI inference cloud (Qorinix), a read-only AI money assistant over Open Banking (LaSpend), a Swiss consumer marketplace (Fixxmi) and an AI-native trading desk. Bridges executives, users, domain experts, technical contributors and vendors, and accelerates discovery and delivery through AI-assisted prototyping and coding agents. Comfortable deep in the technical detail, from LLM, RAG and agentic patterns to inference latency and cost trade-offs, evaluation harnesses and cloud delivery: able to shape requirements, evaluate outputs and direct governed delivery with confidence. MSc Computer Science, AI and Data Science (Merit).

Core Capabilities

AI Product & Transformation

AI strategy and use-case prioritisation
Product vision, MVP definition and roadmaps
Business cases, value hypotheses and adoption planning
Transformation operating models

Product & Delivery Leadership

Product discovery and user needs
Requirements, PRDs, user stories and acceptance criteria
Backlog and release prioritisation
Cross-functional delivery coordination
Stakeholder and executive communication

Responsible AI & Governance

Output and model evaluation
Human-review workflows
Data, privacy, auditability and risk considerations
Provider and vendor evaluation
Quality, cost and operational controls

Financial Services Domain

Corporate finance and M&A workflows
Investment analysis
Quantitative markets and trading
Fintech operations
Risk, compliance and regulated processes

AI Solution Fluency

LLM and RAG concepts
Agentic and workflow-based AI
API, data integration and cloud-delivery concepts
AI-assisted prototyping
Evaluation and quality assurance

How I Deliver

Problem-first use-case selection
Smallest product slice that proves value
Evaluation criteria defined before build
AI-assisted prototyping and coding agents
Quality, cost and latency governed from day one

Professional Experience

Head of AI

Weidmann & Cie. AG

London, UK (Hybrid)December 2024 - Present

Swiss corporate finance advisory firm (M&A and growth advisory) building an applied-AI product portfolio: Qorinix (ultra-low-latency AI inference cloud), LaSpend (read-only AI money assistant over Open Banking), Fixxmi (Swiss consumer marketplace), pricing and insurance comparison workflows, and an AI-native trading desk.

Owns the firm's AI strategy and use-case portfolio across its product lines, prioritising investment against commercial value, risk and regulatory constraints; presents strategy and progress to a finance-led management board
Owns 0-to-1 product direction for Qorinix, the firm's ultra-low-latency AI inference cloud targeting sub-200ms time-to-first-token (p50), 100-200+ tokens per second and sub-10ms cache-hit latency for latency-sensitive workloads such as real-time agents, voice and trading alerts: MVP scope, phased delivery, release gates and go-live criteria across the web, control-plane and AI runtime layers
Defined the platform's evaluation and governance model: latency, quality and cost benchmarks, provider routing and failover, per-call cost ledger, usage-based billing and entitlements, with human-review workflows where risk requires
Shaped consumer-facing propositions: LaSpend, a read-only AI money assistant over regulated Open Banking data (subscription detection, waste scoring and assisted cancellation, never moving customer money, human-in-the-loop for user-triggered actions), and Fixxmi AI lead matching across 12+ service categories (DE-CH/EN) within GDPR and nFADP boundaries
Translates business, operational and regulatory requirements into executable platform plans with senior stakeholders, and directs delivery through technical contributors and AI-assisted development workflows
Commissioned an internal 12-provider inference benchmark to ground model selection, latency and cost decisions; leads a cross-functional team of five across development and financial analysis

AI Engineer & Technical Architect(AI Transformation & Delivery Leadership)

Pacific Cloud Computing Ltd.

Hong Kong & Remote UKDecember 2021 - December 2024

Promoted internally from Senior Software Engineer to lead the firm's AI transformation across SaaS, analytics and market-intelligence products.

Defined and prioritised the firm's AI use-case portfolio and adoption roadmap, moving AI from experiments into production capabilities; oversaw the launch of three multi-tenant SaaS platforms with subscription management, usage analytics and automated billing
Directed delivery of the firm's real-time ML inference capability, a platform that served millions of daily predictions at sub-100ms latency; set requirements, acceptance criteria and evaluation gates
Shaped the firm's enterprise retrieval (RAG) and market-intelligence NLP capabilities, defining evaluation methodology and quality expectations for domain-specific queries
Oversaw the model-improvement cycle (retraining, A/B testing and monitoring) and provider selection across the ML estate
Established AI/ML working practices, ran architecture reviews and mentored an agile development team of 8 engineers
Guided quantitative-markets decision support: cross-asset anomaly detection and price-forecasting research feeding trading workflows

Senior Software Engineer

Pacific Cloud Computing Ltd.

Hong KongJanuary 2015 - November 2021

Delivered an enterprise document management SaaS with real-time collaboration for corporate clients
Delivered e-commerce platform capabilities including multi-currency and crypto-enabled payments; the SEO programme achieved top search rankings
Pioneered the firm's blockchain product initiatives: smart-contract products, NFT marketplace support and DeFi analytics dashboards

Senior Product Manager (Technical) / Web & Content Manager

Groupon.com

Hong KongApril 2013 - December 2014

Led platform modernisation from a monolithic to a service-oriented architecture as technical product manager, coordinating engineering delivery and business stakeholders
Introduced an A/B testing and SEO programme that delivered 150% traffic growth
Ran digital product operations and commercial prioritisation for the marketplace: merchandising, pricing structure and merchant coordination in an Agile environment

Senior Operations Manager

SoManyCall Telecom

Hong KongMarch 2008 - April 2013

Oversaw strategic and operational delivery for custom software solutions in the telecoms sector, achieving a 25% annual growth rate
Led a multi-disciplinary team across the full project lifecycle

Core Technical Competencies

Machine Learning & Deep Learning

Python

PyTorch

Scikit-learn

XGBoost

TensorFlow

Transformers

Decision Trees

LSTM/GRU

Fine-tuning

Feature Engineering

MLOps & Production ML

Automated Retraining

Model Evaluation

A/B Testing

Canary Deployment

Drift Monitoring

Model Serving

ONNX

Triton

Quantisation

Backend, APIs & Microservices

FastAPI

Microservices

REST APIs

Node.js

TypeScript

React

Next.js

Event-Driven

Streaming / SSE

Cloud, DevOps & Infrastructure

AWS

GCP

Cloudflare

Docker

Kubernetes

Terraform

CI/CD

OpenTelemetry

Multi-Region Failover

Data Pipelines & Storage

PostgreSQL

Redis

TimescaleDB

pgvector

Firestore

Cloudflare D1 / R2

Batch Pipelines

Real-Time Pipelines

Generative AI & LLM Systems

RAG

LLM Orchestration

LangChain

Multi-Provider Routing

Prompt Registry

Guardrails

Function Calling

Eval Harness

Research & Selected Projects

Qorinix, Ultra-Fast AI Inference Cloud

Flagship AI

High-speed AI inference cloud for latency-sensitive workloads: real-time agents, voice assistants, trading alerts and conversational AI. Targets sub-200ms TTFT (p50), 100-200+ tokens/sec and sub-10ms cache-hit latency at a fraction of frontier-API cost; internal benchmark targets place it 4-12x faster than leading low-cost providers and 6-12x faster than frontier APIs. Multi-tier model router with provider failover, streaming SSE, per-call cost ledger, usage-based billing and entitlements. Stack: Python, FastAPI, PostgreSQL, Redis, LLM Router, Stripe entitlements, OpenTelemetry.

LLM Arena, 12-Provider Benchmark

LLM Infra

Side-by-side real-time benchmark harness across 12 frontier and fast-inference providers. Streaming token-by-token with live leaderboard, TTFT p50/p95 analytics, tokens/sec, estimated cost-per-task and service-tier selectors. Per-provider model picker, run-count (1-20) for statistical stability and quick TTFT-turbo profiles. Next.js 14, TypeScript, server-side API isolation, streaming-SSE.

LaSpend, AI Money Assistant

Consumer AI

Read-only AI money assistant over regulated UK Open Banking: subscription detection, recurring-spend dashboard, waste scoring and assisted cancellation with verified action receipts. Never moves customer money; human-in-the-loop for all user-triggered actions; PII-minimised inputs and deterministic fallback when models fail. Cloudflare Pages + Functions + D1 architecture, React 19 + Vite + Tailwind frontend.

Fixxmi, AI Swiss Service Marketplace

Consumer AI

Swiss-first B2C service-job lead marketplace with AI-powered intent classification and lead matching across 12+ categories in DE-CH / EN. Pay-per-lead monetisation with Stripe Credit Packs, nFADP / GDPR-compliant data boundaries, Firestore-backed real-time admin dashboard. Next.js 15 static export + Firebase Cloud Functions (Node 20) + Firestore europe-west6.

Audit-Grade AI Governance Loop

Governance

Closed-loop flow applied across every firm AI product: policy check, evidence-pack retrieval, prompt control, model router, action runtime, audit ledger, policy feedback. Every call versioned, costed, support-traceable and replayable. Content-addressed evidence packs, per-product cost attribution, human-in-the-loop escalation for high-risk cases.

Low-Latency Inference Mesh

ML Infra

Distributed ML inference pipeline serving hundreds of predictions per day at <200ms TTFT across classification, sentiment, affordability scoring and anomaly detection. Shard-aware routing, warm-pool autoscaling, ONNX-compiled models with fp16 quantisation, cache-through Redis tiering, multi-region active-active with health-weighted failover. FastAPI, Redis, TimescaleDB, Triton.

LLM-Augmented HFT (MSc Research)

Research

Novel architecture fusing LLM narrative understanding with real-time HFT. Custom transformer with attention heads over microstructure features, ensemble gating and latency-aware inference scheduling that down-routes heavy heads on tight time budgets. AI-assisted development workflow with a three-gate validation framework delivered an 85% reduction in strategy development time. Python, PyTorch, CCXT, Ray Serve, Triton.

Education

MSc Computer Science, AI & Data Science

University of Wolverhampton, UK

2023 - 2025 | Grade: Merit

Dissertation: LLM-Augmented High-Frequency Trading Strategy Development
Modules: Deep Machine Learning, Intelligent Agents, Data Science & Mining, Applying AI, Cloud Computing, Research Methods

Bachelor of Business Administration

Hong Kong University of Science & Technology

1995

Marketing with Information Systems minor

Selected Certifications

MLOps SpecializationGoogle Cloud | 2024

Gemini Certified EducatorGoogle | 2025-2028

CS50x Computer ScienceHarvard | 2024

SFC Licensing Exams (Papers 1/7/8/12)HKSI, Hong Kong | 2021

Open to AI Product, AI Transformation and Head of AI roles across financial services, fintech and applied GenAI.

Full UK right to work | Based near London (Reading) | Hybrid or remote | Languages: English, Cantonese, Mandarin

Book a 30-min chat

Last Updated: July 2026 | Portfolio: www.Philip.pm | References available upon request