Sr. ML Ops Engineer, Forward-Deployed

6sense•San Francisco, CA

1d•Remote

About The Position

This hands-on MLOps engineering role builds and operates production systems and product experiences for customer-tailored predictive models. The engineer works across model training and refresh workflows, APIs, orchestration, data pipelines, production monitoring, and user-facing tools such as model dashboards and training controls. They help turn customer needs into reliable backend systems and usable workflows, validate outcomes, and operate changes safely in production. The role is the first US-based MLOps engineering position, working closely with the MLOps team in India to accelerate customer delivery. Most work is behind the scenes, with customer conversations when technical details matter. The role also turns repeatable customer solutions into reusable product and platform capabilities in collaboration with Product, Data Science, and Platform Engineering. It improves deployment tooling and operational processes, and helps teams launch, understand, refresh, and trust predictive models while reducing fragile manual work.

Requirements

At least 6 years of experience in MLOps, ML platform engineering, backend engineering, full-stack engineering, or related production engineering roles.
Experience owning machine-learning systems in production, including training, deployment, monitoring, retraining, and incident response.
Strong Python and SQL skills.
Experience with APIs, backend services, asynchronous workflows, testing, CI/CD, and observability.
Experience building or materially contributing to internal tools, operational dashboards, or customer-facing product workflows.
Ability to debug across UI, APIs, workflows, data pipelines, model artifacts, and production infrastructure.
Experience safely releasing changes across multiple customers or production environments, including staged rollout, impact checks, and rollback.
Working knowledge of Kubernetes and cloud infrastructure.
Strong communication and judgment when balancing customer urgency, usability, reliability, and long-term maintainability.

Nice To Haves

FastAPI and frontend experience.
Experience with Airflow, Spark, MLflow, model registries, feature pipelines, or workflow orchestration.
Experience with AWS, S3, Hive, distributed compute, and multi-tenant SaaS.
Forward-deployed, consulting, solutions, implementation, or customer-facing engineering experience.
B2B SaaS, go-to-market data, or predictive modeling experience.

Responsibilities

Productionize customer-tailored predictive models, configurations, and scoring workflows with validation, release, monitoring, and rollback paths.
Build and improve UI workflows for training triggers, model refreshes, validation outcomes, model status, and operational troubleshooting.
Enhance model dashboards showing model health, score changes, data quality, refresh status, and production issues.
Build or modify APIs, backend services, data contracts, and orchestration workflows supporting model operations.
Automate model retraining, validation, promotion, and deployment workflows for customer engagements.
Create safeguards for data quality, score distributions, threshold changes, and edge cases before model changes reach production.
Monitor models and services for drift, data shifts, reliability issues, and operational regressions.
Work with Product, Data Science, and Platform Engineering to turn repeatable customer solutions into reusable product and platform capabilities.
Build deployment templates, reusable workflow components, and configuration-driven interfaces that support multiple customers without one-off operational work.
Create self-service or assisted tools for initiating training, monitoring refreshes, reviewing validation results, and safely troubleshooting model issues.
Improve the MLOps platform, including Kubernetes workloads, compute, pipelines, CI/CD, deployment tooling, and observability.
Establish runbooks, release checklists, audit trails, and support processes for model operations.
Identify recurring customer pain points and convert them into product requirements, platform improvements, or durable operational tooling.
Understand the predictive-model lifecycle and improve at least one meaningful deployment, support, or visibility bottleneck within three months.
Bring clear release, monitoring, and support plans to active customer model engagements within three months.
Materially automate common training, refresh, validation, and deployment workflows within six months.
Improve model dashboards or internal workflows to provide clear visibility into model status, health, and refresh progress within six months.
Help customer model changes move to production faster with stronger operational controls within six months.
Turn at least one customer-specific solution into a reusable product or platform capability within twelve months.
Help the team reliably deliver and operate customer-tailored models without fragile manual processes or overnight handoffs within twelve months.