Wonder Group, INC is hiring a Staff Machine Learning Engineer for the Model Platform team. This role provides hands-on technical leadership to define how Java services integrate with ML models, while driving performance, reliability, and cross-team ML architecture from prototype through production.
Role Scope
As a Staff Machine Learning Engineer on the Model Platform team, you will set architectural standards for model calling from Java services and lead work that improves end-to-end performance, stability, and lifecycle management across tier-1 model integrations. You will also translate technical trade-offs for business stakeholders and executives.
Key Responsibilities
- Define the standard architecture for how Java services call ML models, including client libraries, API contracts, latency budgets, and fallbacks.
- Establish standards for load testing, performance benchmarking, and capacity planning for services that call models, and incorporate these practices into the release process.
- Drive performance improvements through profiling, JVM tuning, caching, request batching, and concurrency for highest-traffic model integrations.
- Lead the design and delivery of complex ML solutions spanning multiple teams or systems, making architecture decisions across team boundaries.
- Build reusable ML architectures, pipelines, and deployment patterns that accelerate Applied Science work into production.
- Set standards for ML reliability, scalability, monitoring, and lifecycle management, including model scorecards, reproducible training, safe rollout, and SLOs for tier-1 models.
- Own model architecture and tuning for the most demanding deployments, balancing accuracy with latency, scalability, and cost.
- Run design and code reviews across ML and software engineering teams, challenging technical decisions when supported by evidence.
- Own the platform technical roadmap with engineering leadership, sequencing cross-team work and surfacing risks early.
- Mentor senior and mid-level engineers, participate in hiring, and build team depth so critical systems are not dependent on a single person.
- Explain technical trade-offs in business terms to product and executive stakeholders, and document key decisions and rationale.
Required Qualifications
- BS, MS, or PhD in Computer Science, Engineering, Mathematics, Statistics, or a related field, or equivalent experience.
- 8+ years of professional software engineering experience, with deep expertise building and operating high-traffic backend services in Java.
- 5+ years building and operating ML systems in production, including 3+ years at senior scope setting technical direction across teams or systems.
- Expert Java and object-oriented design skills, including service frameworks (Spring Boot or similar), REST or gRPC APIs, and JVM performance tuning.
- Proven ability to set load testing and performance standards (Gatling, JMeter, or similar) and plan capacity as load changes.
- Experience designing highly available distributed systems (99.99% uptime) with automated testing and monitoring built in from the start.
- Deep production ML architecture expertise, including training pipelines, batch and real-time serving, feature pipelines, model versioning, and large-scale rollback.
- Strong Python and hands-on model development experience with common frameworks (XGBoost/LightGBM, PyTorch, or TensorFlow), including selecting and tuning models under latency and cost constraints.
- Strong data engineering skills in Spark and SQL, including building and debugging streaming or batch pipelines.
- Deep cloud infrastructure experience on AWS or GCP (SageMaker, Vertex AI, or equivalent), plus Docker and Kubernetes.
- Experience with experimentation, including A/B test design, guardrail metrics, and understanding when an offline metric can be misleading.
- Demonstrated technical leadership across ML and software engineering teams, including leading architecture reviews, influencing decisions without direct authority, and mentoring senior engineers.
- Clear communication of model behavior and technical trade-offs to technical and non-technical audiences, including executives.
Technologies
- Java, Spring Boot, REST, gRPC
- JVM performance tuning
- Gatling, JMeter
- Python, XGBoost, LightGBM, PyTorch, TensorFlow
- Spark, SQL
- AWS, GCP, SageMaker, Vertex AI
- Docker, Kubernetes, ONNX Runtime, TensorFlow Java, XGBoost4J
- Cassandra, Kafka, Flink
Benefits
- Competitive compensation package with equity
- 401(k)
- Choice of medical, dental, and vision plans
- Company paid short and long term disability coverage
- Paid time off including flexible time off for exempt employees
- Paid vacation for non-exempt employees
- Paid sick leave in compliance with applicable law
- Paid parental leave
- Discounted meals and exclusive perks across the Wonder family of brands
Hybrid Work Details
The hybrid model requires 3 days per week in the office. Candidates are welcome and encouraged to be in the office up to 5 days per week if it works for them.
Compensation
New York: $208,500 - $216,500 per year. Wonder uses geographic-specific salary structures, and the final offer is based on factors including skills, education, training, credentials, and experience.
Location
New York, NY (hybrid)
Additional Eligibility Policy
Wonder does not sponsor applicants for employment visa status for this role.
Nice-to-Have
- Experience in marketplace, e-commerce, or logistics, including ranking, recommendations, ETA prediction, pricing, or dispatch
- Experience serving models inside or alongside JVM services (ONNX Runtime, TensorFlow Java, XGBoost4J, or similar)
- Building an internal ML platform used by multiple science teams
- Cassandra or other distributed NoSQL databases
- Streaming data systems (Kafka, Flink)
- Leading an ML platform migration between cloud