MLOps Engineer

Cognism is actively seeking an outstanding MLOps Engineer to join our growing Data team. This role is primarily a hands-on engineering and MLOps position, with the individual reporting directly to the Engineering Manager in the Data team. The MLOps at Cognism is entrusted with optimizing and improving the quality of ML services and products. Advising and enforcing best practices within Data Science team, provide tooling and platforms that ultimately results in more reliable, maintainable, scalable and faster Machine Learning workflows. The successful candidate will be at the forefront of our MLOps initiatives, especially during the implementation of our machine learning platform and best practices.

Key Responsibilities:

• Build and manage automation pipelines to operationalize the ML platform, model training and model deployment on AWS

• Design and implement architectures, service and pipelines on the AWS cloud that are secure, reliable, scalable and maintainable

• Operate, maintain and evolve our traditional ML systems in production: gradient-boosted tree models, embedding-based matching, fine-tuned transformer classifiers, and classic NLP pipelines, running on our own serving and data infrastructure as well as some LLM based workflows

• Contributing to the MLOps best practices within the Science and Data team

• Acting as a bridge between AI, Engineering, and DevSecOps for ML deployment, monitoring, and maintenance

• Work closely with Data Scientists to provide tooling and integration of ML models into larger systems and applications

• Monitor and maintain production critical ML services and workloads at scale 

Your Experience:
Required:

• Strong understanding of AWS cloud architecture and services

• Experience deploying and monitoring classical ML models in production on AWS

• Good understanding of modern MLOps best practices

• Good understanding of core Machine Learning fundamentals (classical model training, evaluation, feature engineering not just imited to LLMs)

• Experience serving models in production via a model-serving runtime (e.g. ONNX Runtime, NVIDIA Triton, TorchServe, or similar)

• Experience with vector databases / approximate-nearest-neighbor search (e.g. Milvus, FAISS, pgvector, or similar)

• Good understanding of Data Engineering fundamentals

• Experience with Infrastructure as Code (Terraform, AWS CDK or similar)

• Experience with CI/CD pipelines (GitHub Actions, Circle CI or similar)

• Basic understanding of networking and security practices on cloud

• Experience with containerization (Docker, AWS ECS, Kubernetes, or similar)

• Proficiency reading and writing Python code

• Experience with API deployment frameworks such as FastAPI

• Fluent in English, good communication skills and ability to work in a team

• Enthusiasm in learning and exploring the modern MLOps solutions 

Ideal:

• 3+ years in a MLOps, Machine Learning Engineer or DevOps role

• Ability to design and implement cloud solutions and ability to build MLOps pipelines in AWS

• Good understanding of software development principles, DevOps methodologies

• Experience and understanding of MLOps concepts:
  Experiment Tracking
  Model Registry & Versioning
  Model & Data Drift Monitoring

• Working with GPU based computational frameworks and architectures on cloud (AWS, GCP etc.)

• Knowledge of MLOps and DevOps tools:
  Kubeflow, Metaflow, Airflow or similar
  Visualisation tools – Grafana, QuickSight or similar
  Monitoring tools – Coralogix or GrafanaCloud or similar
  ELK stack (Elasticsearch, Logstash, Kibana)ll,l

• Experience working in big data domains (10M+ scales)

• Experience with streaming and batch-processing frameworks

Bonus:

• Experience with MLOps Platforms (Nvidia Triton, SageMaker, VertexAI, Databricks, or other)

• Knowledge of frameworks such as scikit-learn, Keras, PyTorch, Tensorflow, etc.

• Experience with SQL, NoSQL databases, data lakehouse