> Markdown version of [/jobs/ext/2012417-ai-ml-engineer](https://www.wearedevelopers.com/jobs/ext/2012417-ai-ml-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # AI/ML Engineer - **Company:** Bain & Company - **Location:** Atlanta, GA, United States - **Experience:** Expert - **Salary:** $140,875.0 - $153,750.0 - **Contract:** Permanent contract - **Skills:** Artificial Intelligence, Amazon Web Services, Automated Storage and Retrieval Systems, Code Review, Cron, Information Engineering, Identity and Access Management, Python (Programming Language), Machine Learning, Prometheus, Azure Machine Learning, Management of Software Versions, Large Language Models, Model Validation, Generative AI, Pytest, Information Technology, Low Latency, Code Inspection, Machine Learning Operations, Terraform, Docker, Databricks - **Published:** August 10, 2026 - **Apply:** https://careers.bain.com/jobs/Login?folderId=108807 ## About the Role * Bachelor's degree in Computer Science, Engineering, Machine Learning, Data Science, Statistics, or a related field (or equivalent practical experience). * 6+ years of experience building and operating production ML systems, including model deployment, serving, and post-deployment monitoring. * Demonstrated experience owning the model and prompt lifecycle end-to-end (packaging, registry, promotion, rollout, monitoring, and rollback). * Demonstrated experience building and operating production RAG or retrieval systems end-to-end, from embedding and retrieval through re-ranking and evaluation. * Experience collaborating cross-functionally with Data Science, Data Engineering, and Product teams to ship ML capabilities that solve real user problems. * Strong Python for production ML; code written to production standards (testing, linting, typing). * Experience working in a modern cloud ML platform environment (Databricks or AWS), including managed training / serving and governed model access. * Demonstrated ability to mentor other engineers and raise engineering standards through code review and repository conventions. ML engineering / LLMOps * Strong Python: ML and serving code written to production engineering standards (type hints, Pydantic, pytest, Ruff, mypy strict). * MLflow: experiment tracking, model registry, custom model flavours, promotion workflows, and model serving configuration. * LLMOps tooling: prompt and instruction versioning, model gateways (e.g., Portkey), inference orchestration frameworks (LangChain, LlamaIndex, or equivalent), and response caching. ## Description * Build, deploy, and operate production inference and serving systems for models, embeddings, and re-rankers: request batching, concurrency, and throughput tuning against latency and cost SLAs. * Own the model and prompt lifecycle in MLflow: packaging, model registry governance, promotion workflows, staged rollout behind feature flags, and clean rollback. * Build and maintain LLMOps tooling: prompt and instruction versioning, model-gateway configuration (e.g., Portkey), inference orchestration, and response caching and cost controls. * Build and operate production RAG and retrieval pipelines end-to-end: structure-aware chunking, contextual embedding, hybrid vector plus keyword retrieval, cross-encoder re-ranking, and context assembly. * Design and maintain model and retrieval evaluation frameworks: golden datasets, metric definitions, LLM-as-judge with calibration, regression gates in CI, and production drift monitoring. * Instrument production ML systems with structured logs, OpenTelemetry spans, and Prometheus metrics: token usage, latency percentiles, retrieval hit rates, drift, and hallucination monitoring, with dashboards and alerting. * Collaborate with the Agent / AI squad to serve model and retrieval outputs as structured tool responses consumed by the Agent Gateway; partner with Data Engineers on feature and embedding pipelines. * Drive production ML incident response to resolution; treat deployment, monitoring, and maintenance as part of delivery. Other (20%): * Set and enforce engineering standards for ML and serving code; contribute to repository conventions and raise the bar for production practices. * Mentor mid-level ML Engineers and Data Scientists on production ML and LLMOps practice; conduct thorough code reviews and enforce standards on PRs. * Use AI coding assistants to accelerate pipeline scaffolding, evaluation-harness development, and serving-config authoring; review all generated code against production standards before committing. * Use LLMs to generate first-draft documentation, runbooks, and evaluation reports; validate and refine outputs before publishing., * Model serving and inference optimisation: request batching, concurrency and throughput tuning, latency budgeting, and awareness of quantisation and hardware trade-offs. * RAG pipeline engineering: chunking strategies, contextual embedding, hybrid retrieval, cross-encoder re-ranking, and context assembly. * Vector stores: pgvector, or dedicated vector databases; embedding pipeline design and index tuning at scale. * Model evaluation and monitoring: golden datasets, metric definition, calibration, LLM-as-judge, regression gates in CI, and production drift monitoring. * Feature Store integration: consuming point-in-time correct features (Databricks Feature Store or SageMaker Feature Store) in training and inference. * Docker and Kubernetes: containerising training / inference workloads, writing Job and CronJob manifests, and understanding ephemeral workload patterns. * Infrastructure familiarity: able to provision and review ML-serving infrastructure (IAM roles, model endpoints, GPU / CPU workloads) via Terraform without hand-holding. Generative AI and agentic systems * Builds and maintains inference and retrieval services that feed agent workflows as structured tool responses consumed by the Agent Gateway. * Owns RAG serving quality: designs embedding and retrieval strategies and defines recall / precision and groundedness benchmarks. * Uses LLM-as-judge patterns in evaluation pipelines where qualitative criteria are required. * Uses production feedback and correction signals to design feedback-to-evaluation and feedback-to-training-data loops that close the loop between behaviour and model updates. General * Treats every ML system as a production system from the first commit: tests, observability, a runbook, and an SLA are not optional. * Raises model quality and reliability issues proactively; does not wait for users to report degraded retrieval or drift. * Uses AI tooling to move faster, but reviews all generated code and documentation critically before it enters the codebase. * Strong communication skills; able to explain retrieval-quality, latency, and cost trade-offs to both engineers and non-technical stakeholders. * This role follows a hybrid model, requiring in-office presence at least 1 day per week ## Related Videos - [DevOps for AI: running LLMs in production with Kubernetes and KubeFlow](https://www.wearedevelopers.com/videos/1222-devops-for-ai-running-llms-in-production-with-kubernetes-and-kubeflow) - [pytest: Simple, rapid and fun testing with Python](https://www.wearedevelopers.com/videos/213-pytest-simple-rapid-and-fun-testing-with-python) - [Docker Compose: Rediscovered](https://www.wearedevelopers.com/videos/1978-docker-compose-rediscovered) - [From clicks to cribs - How to find your dream home with web scraping](https://www.wearedevelopers.com/videos/767-from-clicks-to-cribs-how-to-find-your-dream-home-with-web-scraping) - [Automagic Configuration in Python](https://www.wearedevelopers.com/videos/363-automagic-configuration-in-python) - [Docker build without Docker](https://www.wearedevelopers.com/videos/100114-docker-build-without-docker) ## Related Articles - [MLops – Deploying, Maintaining And Evolving Machine Learning Models in Production](https://www.wearedevelopers.com/magazine/115-mlops-deploying-maintaining-and-evolving-machine-learning-models-in-production) - [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it) - [How to Become an AI Engineer](https://www.wearedevelopers.com/magazine/331-how-to-become-an-ai-engineer) - [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development) - [What Are Large Language Models?](https://www.wearedevelopers.com/magazine/304-what-are-large-language-models) - [From Prototype to Production: Build AI Agents with This Free 4-Course Learning Path](https://www.wearedevelopers.com/magazine/655-from-prototype-to-production-build-ai-agents-with-this-free-4-course-learning-path)