MLOPS Engineer 3

Role: MLOPS Engineer 3

Candidates Required: 1

Focus: Production architecture development with scalable & stable infrastructure. Act as a bridge between DS and Dev.

Experience: 4–5 Years

Skillset

Productionization:

Expert in deploying models using Nvidia Triton Inference Server and managing containerized workloads via Docker and Kubernetes (K8s) on EC2.

Feature Stores:

Experience building and maintaining scalable Feature Stores (e.g., Feast, Featureform) to ensure training-serving consistency.

Programming:

Proficiency in Python for ML Deployments, and Kotlin or Java for building robust, scalable backend deployment services. Knowledge of Go for high-performance systems.

Data Systems:

Hands-on experience with Snowflake (as a source), BigQuery, and high-speed databases like Cassandra or Redis for low-latency serving.

DevOps / CI-CD:

Strong command of Bash scripting and CI/CD pipelines (e.g., GitHub Actions, GitLab CI) tailored for ML (Continuous Training pipelines).

Observability:

Setting up and managing monitoring stacks using Grafana and Kibana to track model drift, latency, and system health.

Added Value:

1. Knowledge of Infrastructure as Code (Terraform/CDK) to manage AWS resources (EC2/SageMaker) programmatically.
2. Knowledge of Spot Instances, right-sizing inference servers specific to workload.

Role Expectations

Primary Role: Deployment at scale with Reliability

ML Areas: Sound Knowledge

Deep Learning: Deployment / Quantization

Coding: Python / Java / Go / Bash / SQL

Testing: Load / Stress Testing