AI-powered analysis for financial data
Consulting work building AI products that turn many third-party financial data sources into clear insights, including fine-tuning a specialised model.
- Client
- Confidential client
- My role
- Senior Consultant — AI Products
- When
- 2025 – 2026
- Sector
- Financial data
1 The problem
Clients had access to a large number of third-party financial data sources. More data was not the problem. Turning it into something clear and usable was.
2 What I learned about users
- Scoping with the client showed that what they needed was clarity: the insight, not another feed.
- General-purpose models were not always the right fit. A smaller, specialised model could do the job well while being efficient to run.
3 What I built or changed
- Ran client engagements end to end, from understanding business needs and scoping through to delivery.
- Built AI-powered analysis products that turn many third-party sources into clear, usable insights.
- Designed the analysis algorithms and pattern-recognition logic that find signals and trends across the data.
- Researched and evaluated AI model approaches.
- Fine-tuned a specialised, compute-efficient model on a corpus I helped build, reviewed its outputs and iterated.
4 The result
Products that turn scattered financial data into insight clients can act on, backed by a specialised model that is efficient to run.
- AI product design
- Pattern recognition
- Algorithm design
- Model evaluation
- Fine-tuning
- Data pipelines
- Python
- AWS AI services
- Local LLMs