Middle/Senior ML Developer
Прямой работодатель Burt Intelligence ( burtcorp.com )
Опыт работы от 3 до 5 лет
We are developing a B2B service for advertising inventory forecasting and campaign delivery for large publishers. Its work includes in-depth analysis of available statistics and logs, building forecasting models based on them, and implementing them in a client service. As a result, the service presents an OLAP-like system with API and web interface access, providing a detailed view of expected inventory volumes available for sale.
If you are passionate about data analysis, machine learning, and building innovative solutions for the advertising industry, we encourage you to apply for this exciting opportunity. Join our team and contribute to the development of a cutting-edge B2B service that empowers publishers to make data-driven decisions and maximize their advertising revenue.
You will work in a multicultural team from anywhere in the world:)
What to work on
You will be working across the whole forecasting pipeline. Notable parts are:
- Clickhouse and S3 as a data platform, AWS ECS, Docker to run jobs
- Time series forecasting for thousands of series, daily (Catboost and Etna)
- Clustering and optimization problems to go multidimensional (Python/Scipy and Java)
- Ad network simulation over the unaggregated traffic forecast (Java)
In the following 1-2 years out challenges are: to incorporate external knowledge into the forecasting pipeline, make it more accurate and robust, process and produce more data, improve observability and monitoring, improve performance, reduce new client implementation time, and resolve any other bottlenecks, enabling growth.
Requirements
- Being a good software engineer. 3+ years of production backend or ML development or comparable experience. 1+ year with Python.
- Practical knowledge in ML and statistics. GBDT, Scikit-learn, feature engineering, validation, tuning. Time series experience is a plus. DL and SOTA architectures is not a requirement because honestly we are not there yet.
- Algorithms and data structures, not being scared of them and being able to reason.
- Ability to switch hats and being self sufficient as both data scientist and software engineer.
- Java/Scala/Kotlin experience is a plus.
- MLOps experience is a plus.
- Being ready to develop a product in a customer oriented team, taking responsibility and approach challenges conscientiously.
What you can look forward to with Burt
- Great international & diverse team.
- Join a journey with a growing product and team to build a best-in-class Forecasting solution.
- Cutting-edge technology.
- Collaborative work environment.