Christopher Kelly

Atlas is Arkera's market intelligence platform that uses AI to identify financially relevant information from hundreds of thousands of news articles posted daily in a local language and local news sites. I joined Arkera following the successful completion of an email-based MVP and led product development for the final product. Also, I iterated the product and identified UX pain points and growing daily active users (DAUs).

This regulatory reporting product was created in response to new EU legislation proposed in 2018. Upon publication of the proposals, I identified and validated the need for this product and pitched it to the executive for funding. Funding was successful, and I led the team working in our innovation lab, which built and launched the product. The product was launched globally. Its sales and marketing campaigns were divided into the EMEA, APAC, and American regions. The product uses an underlying blockchain architecture and can be interacted with via the blockchain or an API.

I was commissioned to design a product that would solve a critical problem for AI teams and ensure ongoing compliance in regulated industries. After completing market research, I proceeded to work on problem validation, MVP design, prototype design, go-to-market plan, and business plan. The MVP was pitched to the investment board and is now in development.

I worked with a European banking client to implement a chatbot, taking natural language query expressions from institutional clients and providing a response based on an extensive knowledge database held by the bank. The chatbot was hosted within the existing banking app. Queries were logged to support the ongoing development of the chatbot, and a process was implemented to incorporate new or updated information from the knowledge base and periodic retraining. The project reduced subsequent inbound query volume by 30% over 12 months.

Prior to commencing the project, the client found it hard to maintain a reliable picture of near-real-time financial KPIs. Field incidents, delayed invoices, and other sporadic costs combined with drawn-out month-end processing made costs unpredictable. Volume information was stored separately from contract pricing, and it was challenging to combine them directly to maintain up-to-date revenue forecasts.

The project's goal was to normalize this data in a financial analytics system so the CFO could maintain an accurate picture of finance KPIs from month to month. I led a team of three, including a financial analyst and data engineers. I designed the analytics system and worked with the team to implement it.

Several teams use the system daily for various purposes, including sharing finance KPIs, managing incoming invoices, and ensuring month-end processing accuracy.

A European social media company was managing a large number of different systems. These included several different SaaS products (including Oracle and Salesforce) and several different data products developed independently, either due to the organizational structure or acquisition. The dispersed and independent nature of different data sets made work across them difficult and time-consuming, with no way to combine information from multiple sources.

In this project, I worked with a data warehousing specialist to create a vision for data analytics, build a roadmap to support that vision, and then work with a team to implement it. This included building a reference data management system to normalize different data sources, implementing data schema control, and building out data warehousing pipelines and analytics processes.

As a result, technical and business teams can now continue to work independently with confidence that their data will be stored, normalized, and available for critical analytics functions (including finance, sales, and product).