Founding ML/Data Engineer

What we do:

Astra protects people and brands from the scams, fake shops, impersonation sites, and marketplace abuse that make online commerce less trustworthy.

The threats are criminal, fast, and built to deceive. They steal money or identity of their victims.

Astra helps brands identify, investigate, and takedown online abuse across domains and marketplace ecosystems, turning noisy signals into clear, defensible evidence and prioritized actions.

Who we are:

We’re a small team of highly technical builders, plus a world-class brand protection expert. We’re funded by Icebreaker.vc and a few high profile angels. We have a working product with pilot customers from law firms and brands. The company was founded in November 2025.

Cris, our CEO spent many years at the market leader, Corsearch, and saw an opportunity to set a new standard for this industry.

Position overview:

We are hiring a Founding ML / Data Engineer to own the ML side of Astra’s platform — from raw signals to customer-ready detections and evidence. This role sits at the intersection of backend engineering, data/ML, and product delivery, with real authority to shape how we do ML and what we build next.

You'll work with serious data:

• 420M domain names + metadata
• Hundreds of marketplaces and their millions of products
• Cross-platform signals to find seller similarity and coordinated networks

What you'll do:

• Own ML for detection and prioritization (modeling, evaluation, calibration, drift/error analysis)
• Build data/feature pipelines across domain names + marketplaces (incl. annotation strategies, LLM-assisted labeling where useful)
• Turn outputs into product features: scoring, explanations, drill-downs, and evidence packages that operators trust
• What you’ll focus on will depend on where you believe you can make the greatest impact. A major part of the role is identifying opportunities, running experiments, and shipping improvements end-to-end.

Example project:

• Training on LUMI: deploy repeatable model training + evaluation runs on HPC infrastructure
• Marketplace graph work: similarity between sellers/listings across platforms to identify coordinated abuse

Required qualifications:

• Applied ML experience for classification / scoring (evaluation + error analysis)
• Solid data engineering skills (pipelines, datasets, reproducibility)
• Strong backend/software engineering fundamentals (production-minded, data-heavy systems)
• Comfort with ambiguity: define experiments, make tradeoffs, and ship
• Clear written communication and collaboration

Preferred qualifications:

• Transformers / noisy and imbalanced datasets
• Marketplace data, online fraud, trust & safety, threat intel (or adjacent domains)
• Entity resolution, graphs, clustering (seller networks, infrastructure linkage)
• LLM-assisted extraction/annotation for evidence workflows
• Distributed training / HPC experience (nice to have)
• How the internet works. Domain names, architecture and networking, platforms.

Perks / why join?

• Startup environment: move fast, ship real things, minimal bureaucracy.
• Stock options
• Extreme ownership: high autonomy, high trust, high impact.
• Work on something that matters: protecting brands and consumers from abuse across domains and marketplaces.
• Remote-friendly (work from anywhere in a timezone-compatible way)
• Comprehensive healthcare package
• Lunch benefit