Services
From AI prototypes to reliable, working systems.
Alpina Analytics helps organisations design, develop, test and integrate AI applications. We combine hands-on expertise in AI agents, software engineering and systematic validation with experience in enterprise data, technology consulting, solution engineering and operational workflows.
What we do
AI agents and automation
Build AI agents that work with your data, tools and processes.
- Design agentic workflows, tool orchestration and integrations.
- Connect language models to structured enterprise data, APIs and business applications.
- Implement context retrieval, evidence-backed reasoning and traceability.
- Integrate human review, validation and operational controls.
AI testing and validation
Make AI agent behaviour measurable, testable and easier to maintain.
- Design reference scenarios and end-to-end evaluation strategies.
- Assess answer quality, supporting evidence and tool execution.
- Combine deterministic assertions with semantic validation.
- Automate regression testing and parallel execution.
Our work on TxAgent, including automated testing with Exense Step, demonstrates this approach.
Data engineering and analytics
Turn complex data into reliable inputs for analytics, applications and AI.
- Build data ingestion, transformation and processing pipelines.
- Develop data models, analytical applications and machine learning solutions.
- Integrate cloud and on-premises data systems.
- Process complex financial messaging data, including SWIFT MT and ISO 20022.
AI governance and industrialisation
Bring structure, reliability and control to AI development and deployment.
- Assess AI risks, explainability and traceability requirements.
- Establish validation strategies, governance processes and operational safeguards.
- Support alignment with ISO 42001 and the EU AI Act.
- Prepare AI applications for integration, testing and production deployment.
What Alpina is built on
Enterprise delivery. Hands-on engineering. Deep technical expertise.
Alpina combines experience in technology consulting, solution engineering, enterprise AI, banking operations, software development and data engineering.
Our work spans the full path from problem definition and architecture through implementation, validation, integration and production readiness.
This experience includes:
- designing and delivering data, machine learning and AI solutions in enterprise environments
- building AI agents that connect to business systems, workflows and structured data
- processing and modelling large-scale financial messaging data across SWIFT MT and ISO 20022
- testing, validating and governing AI applications
- supporting cloud, software engineering and industrialisation of production systems
Alpina's approach combines hands-on engineering with the ability to work across business, operations and technology teams.
Experience, technical leadership and collaboration
Experience and technical leadership
Alpina was founded by Pierre Oberholzer, with more than 15 years of experience across technology consulting, solution engineering, data science, enterprise software, banking operations and applied AI.
He spent seven years at a global tier-1 bank, ultimately as Lead Data Scientist in Group Operations, leading cross-functional initiatives involving data science, engineering and banking operations.
Selected experience includes:
- Technology consulting and solution engineering. Translating business and operational problems into technical architectures, prototypes and production-ready solutions.
- Production banking applications. Machine learning to improve sanctions-screening efficiency, meeting a zero-false-negative requirement agreed with Financial Crime Compliance within the evaluated scope, and predictive alerting to help prevent post-trade settlement failures.
- Large-scale financial messaging analytics. Real-time processing and modelling of more than two billion SWIFT MT messages, supporting analytics across banking operations.
- Enterprise AI governance. Development of software incorporating AI regulatory frameworks, knowledge graphs and approaches to trustworthy AI.
- Agentic AI. Architecture, integration and systematic evaluation of AI agents interacting with structured data, tools and enterprise workflows.
At Alpina, this experience has contributed directly to the development of TxFlow, for financial messaging transformation and remediation, and TxAgent, for evidence-backed AI investigations.
Alpina also created PtyParser, an open-source entity recognition library for financial messaging with more than 100,000 downloads since 2020.
Pierre holds a PhD from ETH Zürich and an MSc in Engineering from EPFL. Professional qualifications include Google Cloud Professional Machine Learning Engineer and ISO 42001 Lead Implementer.
Engineering expertise and collaboration
Alpina's products and client work have benefited from collaboration with experienced specialists across software engineering, cloud infrastructure, DevOps, financial messaging and banking technology.
- Software and data engineering. Contributions to financial messaging processing, ISO 20022 schema exploration, data architecture and software development.
- Cloud engineering. Collaboration with experienced Scala and cloud engineers, including through Exense, on TxFlow hardening and preparation for cloud marketplace deployment.
- AI agent engineering. Specialist software engineering and DevOps contributions to TxAgent development and infrastructure.
- Financial messaging expertise. Input from former Swift professionals and independent specialists on messaging standards, product design and market requirements.
- Consulting and delivery. Collaboration with engineering and business specialists where additional expertise or delivery capacity is required.
This model allows Alpina to remain focused while bringing additional specialist capability into an engagement when needed.
How we engage
1. Discover and scope
Understand the business problem, available data and technical environment. Define feasibility, objectives and measurable success criteria.
2. Build and validate
Develop a targeted solution or pilot using representative data. Evaluate its behaviour, performance and operational value with the people who will use it.
3. Integrate and deliver
Support integration into existing systems and workflows, including testing, documentation, deployment and handover according to the agreed scope.
Our focus is practical delivery, technical transparency and solutions that customers can evaluate, integrate and operate within their own environments.
Products and services
Our services can be delivered independently or alongside Alpina's software products.
- TxFlow structures, transforms and remediates financial messaging data.
- TxAgent investigates payment exceptions using evidence-backed AI.
Both products reflect the engineering and domain expertise behind our consulting work. Neither is a prerequisite for an engagement.
Discuss your project
Have an AI, data engineering or automation challenge?
Whether you are developing an AI agent, establishing a testing strategy or integrating complex data into operational workflows, we can help assess the challenge and define a practical path to delivery.