About Us: Attentive.ai is a fast-growing vertical SaaS start-up, funded by Peak XV (Surge), InfoEdge, Insight Partners, Vertex Ventures, and Tenacity Ventures that provides innovative software solutions for the landscape, paving & construction industries in the United States. Our mission is to help businesses in this space improve their operations and grow their revenue through our simple & easy-to-use software platforms.
Position Description: As an SDE-II AI Engineer, you will sit at the intersection of our AI Research and Engineering teams, owning the path that takes computer vision, NLP, and multi-modal models from research prototypes to reliable, scalable production systems. You will build and operate the infrastructure, pipelines, and tooling that let our models run efficiently in production, powering automated construction take-off and estimation from blueprints, drawings, and PDF documents.
In this role, you will work closely with Research Engineers and Backend Engineers to close the gap between experimentation and deployment, designing training and inference pipelines, setting up experiment tracking and model versioning, and building the monitoring and observability that keep our AI systems accurate and dependable at scale.
Roles & Responsibilities:
- Own the end-to-end MLOps lifecycle, from model packaging and CI/CD to deployment, monitoring, and rollback for computer vision, NLP, and multi-modal models.
- Design and maintain scalable training and inference pipelines for large datasets and models, optimizing for cost, latency, and throughput.
- Build and manage containerized deployment infrastructure (Docker, Kubernetes) for hosted deep learning and geoprocessing services.
- Set up and maintain experiment tracking, model registry, and versioning systems to ensure reproducibility across the research-to-production lifecycle.
- Implement model monitoring and observability — drift detection, performance degradation alerts, logging, and dashboards, for models running in production.
- Apply model optimization techniques (quantization, pruning, knowledge distillation) to improve inference efficiency in production.
- Collaborate with Research Engineers, Backend Engineers, and Product teams to translate research ideas into deployable, production-ready services.
- Develop and maintain infrastructure-as-code, monitoring, and logging for all deployed ML/AI software.
- Evaluate, profile, and continuously improve the reliability, scalability, and cost-efficiency of existing ML systems.
- Stay current with evolving MLOps tooling and best practices and evaluate applicability to construction industry challenges.
Skills & Requirements
- 3+ years of experience in MLOps, ML infrastructure, or applied AI/ML engineering, with exposure to Computer Vision or NLP systems.
- Hands-on experience with workflow orchestration frameworks (preferably Temporal) for building reliable, fault-tolerant, long-running distributed workflows.
- Strong proficiency in Python and hands-on experience with ML frameworks such as PyTorch, TensorFlow, OpenCV, or HuggingFace Transformers.
- Hands-on experience with Docker, Kubernetes, and containerized ML deployment pipelines in production environments.
- Experience building and maintaining CI/CD pipelines for ML systems (e.g., Jenkins, GitHub Actions, GitLab CI).
- Working knowledge of experiment tracking and model registry tools (e.g., MLflow, Weights & Biases, DVC).
- Experience with cloud environments (GCP, AWS, or Azure) and infrastructure-as-code practices.
- Familiarity with messaging and data processing pipelines (e.g., Apache Kafka, RabbitMQ) and distributed web services.
- Understanding of model optimization techniques such as quantization, pruning, and knowledge distillation - Good to have.
- Experience with version control systems (e.g., Git) and project tracking tools (e.g., JIRA).
- Familiarity with PostGIS or other geo-databases and Geographic Information Systems - Good to have.
- Strong analytical and problem-solving skills with a passion for building reliable, scalable systems; comfortable working in a fast-paced, agile, startup-like environment.