About Yf
Data Scientist specialising in applied AI and generative systems, with a track record of delivering production-grade ML and LLM solutions across major UK publishing and news organisations. I design and build cloud-native AI platforms on GCP — from RAG pipelines and agentic systems to semantic search and speech recognition — that automate workflows, surface insights, and unlock new revenue streams. Known for bridging technical delivery and business strategy through clear analytical storytelling and close collaboration with editorial, commercial, and senior leadership stakeholders.
English
Native or bilingual
Cantonese
Native or bilingual
Chinese
Native or bilingual
Remote only
Primarily works remotely
Experience
- Major International Publishing HouseData ScientistDIGITAL AND ITJanuary 2026 - Today (5 months)London, United Kingdom
- Architecting a GCP-native data and AI platform (BigQuery, Vertex AI, Google ADK) to unify metadata, sales, and content for AI-enhanced BI and backlist discovery
- Designing agentic AI systems with Google ADK for trend detection, sales attribution, and semantic search over enterprise data
- Building production RAG pipelines combining BigQuery vector search, Vertex AI embeddings, and hybrid retrieval approaches (LightRAG, BYOC on Vertex AI RAG Engine)
- Engineering ingestion pipelines for industry-standard XML feeds (ONIX) and integrating heterogeneous systems across ERP, CRM, and legacy reporting datamarts
- Leading POC delivery end-to-end, from architecture through stakeholder demos to senior commercial leadership
- Developing fraud detection pipelines for supplier financial data with multi-rule scoring across ~15K records
- Production GCP hardening: IAM least-privilege scoping, service account architecture, and BigQuery cost/performance optimisation
- Major UK News PublisherAssociate Data ScientistDIGITAL AND ITJanuary 2024 - January 2026 (2 years)London, United Kingdom
- Built and productionised a PyTorch/WhisperX speech recognition and diarisation system with LLM summarisation, scaling to 1K+ monthly active users across UK, US, and Australian business units
- Automated extraction from 20K historical index cards using LLM-based parsing on Vertex AI / GCP, generating a searchable archive projected to drive £1.2M in licensing revenue
- Developed an embedding-driven semantic search prototype with vector database retrieval, auto-generating event timelines for journalists during article publication workflows
- Partnered with editorial teams to build an LLM-powered tool converting digital articles into print layouts, validating a workflow now progressing to production
- Led an end-to-end ML project on BigQuery benchmarking article performance, from data cleaning to model deployment via CI/CD, supporting newsroom content curation
- Provided data science advisory to leadership on early-stage AI initiatives, defining and validating high-impact use cases for newsroom and archive automation
- Worked across PyTorch, LangChain, Vertex AI, BigQuery, and CircleCI to deliver production-grade ML and LLM solutions integrated with editorial workflows
Recommendations
Be the first to recommend Yf
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Certifications
- AWS Certified Solutions Architect – AssociateAWS2025