SEROKELL

Serokell was hired by Wisk, an air mobility company, to develop a fully automated, two-tier machine learning prediction system for dynamic air ticket pricing. The team at Serokell designed a price modifier based on historical data, implemented a random forest as the primary model, and switched the evaluation procedure from MAPE metric to SMAPE, resulting in an error correction improvement of approximately 10-15%. They also reduced service costs and increased customer ROI through dynamic price updates based on ML modifiers. The project was successful and Serokell was commended for its high ratings, good value for cost, and alignment with company values.

For a Spectroscopy Tools Company, Serokell was tasked with the ground-up rewrite of core instrument control software for a new platform. The original software, written in Python, had evolved from prototype quality code, and was now around 10 years old. Serokell rewrote the software using Rust, and also assisted in the design of the UI/UX. The project is ongoing, but the company has praised Serokell for their high technical expertise, great culture fit, and good value for cost.

Agecroft Partners, an investment consulting company, hired Serokell to develop an advertising platform that automates the targeting process in ad campaigns through sophisticated keyword analysis. The ML model was trained to break down a user's query into organized segments of structured language and identify the key elements. As a result, users marketers can provide simple wishes about PPC campaign details; all necessary information is gathered directly from a query expressed in everyday conversational English. Serokell received high ratings, was a great culture fit, and was referred to the client, leading to a successful project completion.