We are seeking a highly skilled and experienced Staff Software Engineer to join our Communication Products team. As a Staff Engineer, you will be responsible for leading the technical vision for ML-powered messaging features, architecting and delivering intelligent capabilities end-to-end, and partnering deeply with ML and product teams.

The Difference You Will Make:

As a Staff Engineer on the team, you will define and drive the technical strategy for integrating ML capabilities into Airbnb's messaging products, including smart replies, message classification, content moderation, translation, and conversational assistance. You will also own the full lifecycle of ML-powered features: from prototyping and experimentation through launch, monitoring, and iteration.

A Typical Day:

• Design, build, and operate the systems that serve ML models within the messaging stack, with a focus on latency, reliability, and scalability
• Write and review technical designs that solve large, open-ended problems at the intersection of ML and product engineering without clearly-known solutions
• Partner with ML, data science, and product teams to identify high-value opportunities, establish evaluation criteria, and close the gap between offline model performance and production impact
• Collaborate with other engineers and cross-functional partners across Messaging, Trust & Safety, Localization, and Platform organizations to align on long-term technical solutions
• Mentor, guide, advocate, and support the career growth of individual contributors
• Establish engineering standards for ML integration across the messaging surface, including feature flagging, A/B testing, observability, and graceful degradation

Your Expertise:

• 9+ years of relevant engineering hands-on work experience
• Bachelors, Masters, or PhD in CS or related field
• Demonstrated experience building and shipping ML-powered product features in production environments, including model serving, feature pipelines, online/offline evaluation, and monitoring
• Exceptional architecture abilities and experience with architectural patterns of large, high-scale applications
• Familiarity with NLP/NLU techniques and large language models, particularly as applied to messaging, conversational AI, or content understanding
• Shipped several large-scale projects with multiple dependencies across teams, specifically at the intersection of ML infrastructure and product engineering
• Technical leadership and strong communication skills with the ability to translate between ML research, product goals, and engineering execution
• Experience operating distributed, real-time systems at scale with high reliability requirements
• Experience with real-time messaging systems or event-driven architectures
• Familiarity with ML infrastructure at scale (e.g., feature stores, model registries, online inference platforms)
• Prior work on trust & safety, content moderation, or internationalization in a messaging context
• Experience with LLM-based product features, including prompt engineering, retrieval-augmented generation, or fine-tuning

How We'll Take Care of You:

Our job titles may span more than one career level. The actual base pay is dependent upon many factors, such as: training, transferable skills, work experience, business needs and market demands. The base pay range is subject to change and may be modified in the future. This role may also be eligible for bonus, equity, benefits, and Employee Travel Credits.

Pay Range: $204,000-$255,000 USD

XML job scraping automation by YubHub

On this team, you will build and operate the ML infrastructure and product services that enable trust and intelligence across Plaid's network. You'll own feature engineering, offline training and batch scoring, online feature serving, and real-time inference so model outputs directly power partner-facing fraud & trust products and bank intelligence features.

Responsibilities

• Embed model inference into Network Enablement product flows and decision logic (APIs, feature flags, backend flows).
• Define and instrument product + ML success metrics (fraud reduction, retention lift, false positives, downstream impact).
• Design and run experiments and rollout plans (backtesting, shadow scoring, A/B tests, feature-flagged releases) to validate product hypotheses.
• Build and operate offline training pipelines and production batch scoring for bank intelligence products.
• Ship and maintain online feature serving and low-latency model inference endpoints for real-time partner/bank scoring.
• Implement model CI/CD, model/version registry, and safe rollout/rollback strategies.
• Monitor model/data health: drift/regression detection, model-quality dashboards, alerts, and SLOs targeted to partner product needs.
• Ensure offline and online parity, data lineage, and automated validation / data contracts to reduce regressions.
• Optimize inference performance and cost for real-time scoring (batching, caching, runtime selection).
• Ensure fairness, explainability and PII-aware handling for partner-facing ML features; maintain auditability for compliance.
• Partner with platform and cross-functional teams to scale the ML/data foundation (graph features, sequence embeddings, unified pipelines).
• Mentor engineers and document team standards for ML productization and operations.

Qualifications

• Must-haves:
• Strong software engineering skills including systems design, APIs, and building reliable backend services (Go or Python preferred).
• Production experience with batch and streaming data pipelines and orchestration tools such as Airflow or Spark.
• Experience building or operating real-time scoring and online feature-serving systems, including feature stores and low-latency model inference.
• Experience integrating model outputs into product flows (APIs, feature flags) and measuring impact through experiments and product metrics.
• Experience with model lifecycle and operations: model registries, CI/CD for models, reproducible training, offline & online parity, monitoring and incident response.
• Nice to have:
• Experience in fraud, risk, or marketing intelligence domains.
• Experience with feature-store products (Tecton / Chronon / Feast / internal) and unified pipelines.
• Experience with graph frameworks, graph feature engineering, or sequence embeddings.
• Experience optimizing inference at scale (Triton/ONNX/quantization, batching, caching).

Additional Information

Our mission at Plaid is to unlock financial freedom for everyone. To support that mission, we seek to build a diverse team of driven individuals who care deeply about making the financial ecosystem more equitable.

XML job scraping automation by YubHub