About Sifflet
We are building the world’s best data observability platform to help companies excel at data-driven decision making.
Today, half of a data team’s time is spent troubleshooting data quality issues. Sifflet is putting an end to that. Our solution allows data engineers and data consumers to visualize how data flows between their services, define data quality checks, and quickly find the root cause of any data anomaly.
About the job
The monitoring team implements the foundational capabilities of Sifflet: detecting data quality issues across a wide range of data warehouses and databases.
Sifflet’s monitoring capabilities rely heavily on machine learning (ML) techniques. Most advanced data quality checks are based on time series forecasting models that detect unexpected distribution changes while accounting for seasonality and one-off events. Additionally, ML-based features are present throughout our product, be it for intelligent alert grouping, automated incident description, or automated monitor suggestions.
As a machine learning engineer on the monitoring team, you will:
- Build automated data profiling systems that learn normal data patterns and detect deviations.
- Deploy time series forecasting models that understand data seasonality and business cycles.
- Create intelligent alerting systems that reduce noise through ML-powered incident correlation.
- Implement generative AI workflows across the product, such as enabling users to describe their monitoring needs in natural language.
- Contribute to product decisions and identify areas where adding ML/AI-based features can solve customer pain points.
As we’re a small team, you will be expected to design, implement, deploy and maintain your projects in production, and integrate them with other services. Thus, this role includes a significant software engineering component.
Some projects you could be working on
- Automated monitor recommendations based on data profiling metrics.
- Automated root cause analysis of any data quality incident, building upon the many sources of metadata Sifflet collects (table lineage, query history, past monitor failures…).
Our stack
- The monitoring engine is built with Python 3 and its large data/ML ecosystem (notably PyTorch).
- The web API is written in (modern) Java with Spring Boot 3, the web frontend is a VueJS application written in Typescript. You may occasionally need to make minor changes to this code base.
- Infrastructure: Kubernetes (AWS EKS clusters), MySQL (on AWS RDS), Temporal for job orchestration
- Plus a few supporting services: Gitlab CI, Prometheus/Loki/Grafana, Sentry…
While not directly part of our stack, expect to gain a lot of knowledge on many products in the modern data ecosystem. The subtleties of BigQuery or Snowflake will soon be very familiar to you.
Are we the company you’re looking for?
- We have offices in Paris, but we’re very remote friendly - several team members are fully remote.
- We offer competitive salary and company equity.
- We have experts on many topics, so there’s always someone to help. We also have tech talks where everyone can discuss a cool project or technology.
- We’re constantly exposed to the intricacies of the modern data ecosystem - you’ll become very knowledgeable about data engineering and the modern data stack, and about how data is used in enterprises.
All written communication at Sifflet is in English, but the engineering team routinely uses French, so some level of fluency in French is required.
None of the people who joined Sifflet perfectly matched the described requirements for the role. If you’re interested in this position but don’t tick all the boxes above, feel free to apply anyway!- A first call with either Benoît (Head of Engineering) or Pierre (Monitoring Team Lead) (30 minutes)
- One coding interview and one system design interview (1h30 each)
- One call with a product manager (30 minutes)
At this point we generally know if we want to extend an offer, but we’re happy to organize additional sessions so you can better know the team and the company.
