Founding Senior ML Engineer at Afori

on-site · full_time · Visa sponsorship

Apply for this role at Afori

Responsibilities:
- Design and own the continuous learning (learning flywheel) architecture for agent swarms: capture agent outcomes, expert signals and case feedback to drive iterative model improvements.
- Build and operate fine-tuning and distillation pipelines for German insurance NLP tasks: dataset curation, training, validation, and staged rollouts.
- Define and implement evaluation standards: gold sets, benchmarks, abstention/faithfulness checks, and production eval gates.
- Develop model serving and cost-aware inference strategies to run at scale across thousands of real Cases in a regulated environment.
- Create cross-sell intelligence models that score opportunities from APDB/account data and feed actionable signals into broker workflows.
- Collaborate with product, infra, legal and ops to ensure compliance, observability and safe deployments; monitor production performance and iterate.

Requirements:
- Proven experience building and shipping ML systems in production, including LLM fine-tuning and model distillation workflows.
- Hands-on expertise designing evaluation methodologies (gold sets, benchmarks, abstention checks) and enforcing eval gates for production.
- Strong ML engineering fundamentals: training pipelines, serving infrastructure, monitoring and cost-aware inference at scale.
- Track record owning ambiguous end-to-end problems in early-stage or fast-moving product contexts; bias for finishing and operational rigor.
- Experience with NLP in regulated domains, agentic systems, continuous learning loops, or German-language models is a strong plus.
- Fluent Python and familiarity with model/tooling ecosystem (open-weight and proprietary models); good trade-off judgment between performance, cost and safety.