Uprevo protects one-click subscription flows used by MNOs and VAS providers, where header enrichment identifies the end user’s MSISDN for immediate activation—an area highly targeted by automated fraud and hidden WebViews (fraud rates can exceed 95%). You’ll build advanced detections that operate in real time and support asynchronous post-processing cancellation of fraudulent activations.
What you’ll do
• Design and ship algorithms that learn patterns in activation-click coordinates (e.g., repeated pixel hits, clustered distributions) to classify traffic and either block directly or score adaptively.
• Evolve current honeypots (hidden inputs/fields) and introduce tiny “visible” decoy buttons to trap bots/programmatic clicks reliably.
• Extend the dual-pixel setup (header-enriched + non-header-enriched) with multi-pixel strategies and stronger validation logic to detect data mismatches and evasions.
• Improve server/client measures across the secure activation flow; optimize verification endpoints and support async cancellation windows after deeper analysis.
• Build within a microservices ecosystem (6–7 Spring Boot/Java services), dockerized and running on on-prem infrastructure; evolve the JS landing page and React admin portal; work with PostgreSQL, Cassandra, and Redis.
• Contribute tests, keep CI/CD green, and help close doc gaps across a large codebase.
• Partner with the current Uprevo devs and cross-functional stakeholders to prioritize fraud measures that matter most.
Qualifications
• Proven experience designing fraud or abuse detections (rules, heuristics, and/or ML) for web traffic.
• Strong backend engineering with Java / Spring Boot and microservices; hands-on Docker.
• Comfortable with PostgreSQL, Cassandra, Redis and high-throughput data validation/aggregation.
• Frontend familiarity to instrument/ship detections on a JS landing page and React admin.
• Experience building honeypots, tracking pixels, or click-path/coordinates analytics.
• Good test hygiene and CI/CD mindset for large codebases.
Nice to have
• ML for pattern recognition (e.g., clustering/DBSCAN for coordinate patterns, anomaly scoring).
• Telecom / header-enrichment know-how (MSISDN flows), bot/malware detection, or anti-ad-fraud background.• Experience operating on-prem deployments at scale.