Senior MLOps Engineer

Client Overview: Our client is an Azerbaijani telecommunications company, the largest mobile network operator in Azerbaijan. The main products are: Fixed telephony, Mobile telephony, Internet services, Wireless broadband, and Value-added services. Project Objectives: The primary goal is to accelerate the client’s Data & AI initiatives via a secure, hybrid cloud foundation on AWS while systematically modernizing the IT estate as part of the cloud migration. Key Project Objectives include: Cloud Foundation & Landing Zone: Deploy target hybrid network architectures, establishing a secure Landing Zone and hybrid Data/AI platforms on AWS. Security, Compliance & Governance: Operationalize on-prem tokenization (achieving zero raw PII in the cloud), resolve policy blockers to include AWS in the ISMS, and establish a Cloud Center of Excellence (CCoE) to govern Cloud adoption. AI Chatbot & Voicebot Design & Implementation: Develop and operationalize a flagship Customer Care Chatbot and Voicebot as the first hybrid-setup consumer. Responsibilities: Build, operationalize, and automate end-to-end MLOps pipelines using Amazon SageMaker Pipelines and MLflow for experiment tracking, model versioning, and registry lifecycle management. Design, deploy, and manage production SageMaker inference endpoints (real-time, serverless, and batch) and Amazon Bedrock API integrations for LLM/SLM deployment with cost controls and latency optimization (Bedrock API Gatekeeper). Implement AgentOps / LLMOps frameworks (AgentCore, Bedrock Guardrails, Promptfoo) to manage multi-agent orchestration, prompt evaluation, safety guardrails, and RAG retrieval pipelines. Operationalize real-time STT / TTS (Speech-to-Text / Text-to-Speech) voicebot pipelines and low-latency speech inference on hybrid/cloud GPU node pools for the flagship Customer Care Voicebot. Optimize specialized GPU node pools (NVIDIA A100/L40S / EC2 GPU instance types) for Azerbaijani SLM/LLM model training, fine-tuning, and scalable inference workloads. Establish automated CI/CD for Machine Learning using GitLab CI/CD pipelines and Infrastructure-as-Code ( Terraform or AWS CDK ) to enforce security-gated MLOps promotion workflows (from SageMaker Canvas/Sandbox to production). Integrate data de-identification, Format Preserving Encryption (FPE) , and tokenization wrappers into ML data pipelines to ensure zero raw PII enters AWS cloud environments during model training and inference. Set up telemetry, performance monitoring, model drift detection, and cost anomaly alerting for AI/ML workloads using Amazon CloudWatch , Splunk , and FinOps spend control frameworks. Collaborate with Data Engineering, AI Architects, and Cloud Teams to integrate vector storage/retrieval (RAG), Apache Spark/EMR-on-EKS runtimes, and local tokenization databases. Author technical MLOps runbooks, model deployment procedures, governance documentation, and disaster recovery playbooks. Requirements: 4+ years of hands-on experience in MLOps, DataOps, or Platform Engineering with a primary focus on enterprise Amazon SageMaker (Pipelines, Feature Store, Model Registry, Endpoints). Proven experience deploying and operating Generative AI, LLM/SLM models, and Amazon Bedrock services alongside agentic frameworks and RAG pipelines. Hands-on expertise with MLflow for experiment tracking, model registry, and lifecycle management. Solid experience in GPU optimization and orchestration (NVIDIA A100/L40S, AWS EC2 GPU instances) for model training, fine-tuning, and low-latency real-time inference (STT/TTS voice pipelines). Proficient in building CI/CD for Machine Learning (GitLab CI/CD, GitHub Actions) and Infrastructure-as-Code ( Terraform or AWS CDK ). Practical knowledge of LLMOps / AgentOps tools and methodologies (AgentCore, prompt evaluations, Bedrock Guardrails, vector databases for RAG). Strong understanding of data security, privacy, and tokenization (FPE, handling sensitive/PII data within ML pipelines). Proficient in Python, PySpark, Docker, and Kubernetes/EKS fundamentals for containerized ML workloads. Nice-to-Have Skills: AWS Certified Machine Learning – Specialty certification. AWS Certified Solutions Architect – Associate/Professional or AWS Certified DevOps Engineer – Professional . Experience in telecom domain AI/ML applications, low-latency real-time voice/chat processing (ASR/TTS), or hybrid cloud data sovereignty architectures. Experience with EMR-on-EKS, Starburst/Athena, or Apache Iceberg data lake integrations. Soft Skills & Team Fit: Strong critical thinking, problem-solving, and analytical skills. Excellent communication and collaboration skills to work closely with cross-functional teams (Data Engineering, AI/GenAI Engineers, Security, Cloud/Infrastructure). Results-oriented, proactive mindset with strong ownership of deliverables within an Agile / Scrum framework. Upper-Intermediate+ English level (written and spoken). What we propose: Opportunity to lead critical, high-impact Data & AI platform delivery for a major telecommunications operator. Hands-on work with modern MLOps and GenAI stack (Amazon SageMaker, Amazon Bedrock, MLflow, AgentCore, STT/TTS voicebot pipelines). Flexible remote work options with structured, predictable collaboration within a well-balanced team. We offer*: Flexible working format - remote, office-based or flexible A competitive salary and good compensation package Personalized career growth Professional development tools (mentorship program, tech talks and trainings, centers of excellence, and more) Active tech communities with regular knowledge sharing Education reimbursement Memorable anniversary presents Corporate events and team buildings Other location-specific benefits *not applicable for freelancers