Our Work: Case Studies in Data & Risk
Production implementations at the intersection of finance, risk, and technology
From enterprise data warehouses for major financial institutions to pioneering blockchain credit scoring models, our technical implementations are designed to deliver measurable, real-world impact at scale. Explore some examples of our work below.
The Challenge
A major financial institution was operating with over 10 disconnected systems and 100+ manual Excel reports, resulting in week-old data and conflicting metrics across departments.
.jpg)
Our Solution
We designed and implemented a comprehensive enterprise data warehouse from the ground up, transforming their entire data ecosystem.
Architecture: A robust star schema with Slowly Changing Dimensions (SCDs) to support both historical analysis and real-time reporting.
ETL/ELT: 50+ Alteryx workflows to automate data extraction, transformation, and loading with complete error handling.
Analytics: A Power BI semantic layer with 150+ calculated measures to ensure metric consistency.
Governance: Implemented row-level security, data classification, and complete audit trails for regulatory compliance.
Impact
80%
reduction in reporting time (from 5 days to same-day).
100%
audit-ready compliance for data governance requirements.
A single source of truth
that eliminated conflicting metrics.
Enterprise Data Warehouse
A Complete Digital Transformation
The Challenge
A digital lending platform needed to move from overnight batch processing to real-time analytics while maintaining complete audit trails for regulatory compliance.
.jpg)
Our Solution
We architected an event-sourced system using a modern data stack, enabling instant insights with perfect historical reconstruction.
Architecture: Implemented an immutable event log in ClickHouse capable of handling over 100,000 events per hour.
Transformation: Built 150+ dbt models with incremental processing to handle late-arriving events.
Analytics: Developed self-service analytics in Metabase with sub-second query performance.
Auditability: Enabled point-in-time reconstruction of any historical state for regulatory reviews.
Impact
60%
average decision-making time.
100%
event capture with zero data loss.
Real-time visibility
Moved from a 24-hour lag to real-time visibility into operations.
Event-Sourced Analytics
Real-Time Data at Scale
The Challenge
A national credit bureau needed to implement a comprehensive data governance and privacy framework to comply with new regulations (analogous to GDPR/PDPL) while processing sensitive financial data for millions of customers.
.jpg)
Our Solution
We designed and implemented a complete data privacy and governance framework that became the regional benchmark for compliance.
Privacy by Design: Built data minimisation, purpose limitation, and consent management into every system.
Access Control: Deployed attribute-based access control (ABAC) with dynamic masking based on user context and data sensitivity.
Data Subject Rights: Engineered cascading deletion and access request mechanisms across distributed systems.
Impact
100%
regulatory compliance achieved in the first audit.
48h
response time for data subject requests (where regulations require 30 days).
95%
reduction in manual compliance tasks.
Data Governance & Privacy Excellence
The Challenge
A national credit bureau needed to move beyond basic data provision to offer a predictive scoring model but lacked the in-house expertise to develop and deploy ML at scale.
.jpg)
Our Solution
We built a complete ML pipeline from data preparation to production deployment, creating a bureau score that became the new industry standard.
Data Integration: Combined traditional credit data with alternative sources including telecom, utilities, and trade payment behaviour.
Feature Engineering: Developed over 200 predictive features, including payment velocity, stability indices, and utilisation patterns.
Model Development: Tested multiple algorithms (XGBoost, LightGBM, Logistic Regression) with rigorous validation and selected the optimal challenger.
Monitoring: Implemented a full monitoring framework, including drift detection (PSI), performance tracking, and automated retraining triggers.
Impact
0.85 Gini coefficient
Achieved over the industry benchmark of 0.70-0.75.
100%
adoption across more than 50 financial institutions.
Sub-second scoring
Enabled sub-second scoring for real-time decisioning.
Machine Learning Bureau Score
From Concept to Production
The Challenge
Corporate lending teams required sophisticated yet accessible risk assessment tools that could handle complex financial analysis while remaining auditable and transparent for credit committees.
.jpg)
Our Solution
We developed comprehensive Excel-based models that balanced analytical sophistication with usability, becoming the standard for decisions on credits from $1M to $500M.
Dynamic Analysis:
Built models that automatically adjust to different accounting standards and company structures.
Scenario Modelling: Integrated Monte Carlo simulations for stress testing with over 10,000 iterations.
Automated Reporting: Enabled one-click creation of credit memorandums with all supporting analysis.
Impact
60%
reduction in analysis time per credit application.
Standardised assessments
across all analysts, improving consistency.
Zero spreadsheet errors
achieved through robust validation rules.
Financial Risk Models
Institutional-Grade Excel Solutions
The Challenge
DeFi protocols needed a way to assess creditworthiness, but traditional scoring models were unable to evaluate on-chain behaviour and wallet history.
.jpg)
Our Solution
We pioneered one of the first production blockchain credit scoring systems, combining on-chain transaction analysis with optional off-chain data sources.
Multi-Chain Data: Integrated data from Ethereum, BSC, and Polygon, including major DeFi protocols like Aave and Compound.
Blockchain-Native Features: Developed novel features for wallet age, protocol experience, liquidation history, and portfolio composition.
Privacy Preservation: Implemented zero-knowledge proof concepts for sensitive data.
Impact
50,000
ccored wallets within the first year.
First-to-market
with a production blockchain credit scoring API.
New market
for under-collateralised lending in DeFi.
Blockchain Credit Scoring
Bridging DeFi and TradFi
The Challenge
A crypto hedge fund needed to compete with MEV bots and high-frequency traders while managing risk across highly volatile markets.
.jpg)
Our Solution
We built a multi-agent AI system that analyses market data, generates signals, and executes trades with sophisticated, automated risk management.
Architecture: Orchestrated specialised agents for data collection, signal generation, risk assessment, and execution.
MEVProtection:
Implemented Flashbots integration and commit-reveal schemes to prevent front-running.
Risk Management:
Built-in algorithms for position sizing, stop-loss, and portfolio rebalancing.
Impact
2.8
Achieved Sharpe ratio in live trading.
12%
maximum drawdown maintained during significant market stress.
$5M AUM
Grew within the first 6 months.
AI Trading Agents
Autonomous Market Intelligence
Our technical stack
While we are technology-agnostic, we have deep expertise in a core set of modern, scalable tools.
Data Platforms
ClickHouse, Alteryx, Dataiku
Engineering
Python, SQL, dbt, Elixir
ML/AI
PyTorch, TensorFlow, XGBoost, LangChain, OpenAI, Hugging Face
Visualisation
Power BI, Metabase, Tableau
Cloud
AWS, GCP
