Jobs at Robots & Pencils

Lead the architecture and delivery of production AI and GenAI systems, including agentic workflows, RAG pipelines, evaluation, observability, reliability patterns, and AWS cloud services. Build backend and serverless applications in Python and Node.js, support frontend integrations, mentor engineers, and own features through deployment and troubleshooting. The role requires weekly onsite client delivery in the United States and Canadian work authorization with eligibility for U.S. TN status.

Build and deploy large-scale LLM-based recommender systems from raw customer and catalog data through production inference. Responsibilities include feature and context engineering, distributed training on SageMaker HyperPod, LoRA/QLoRA fine-tuning, reinforcement learning, deep learning architecture, ranking, inference optimization, evaluation, privacy protection, and stakeholder communication with Expedia and AWS science teams.

Lead the design and implementation of a production-grade evaluation framework for planner-executor AI agents. Build simulation, experiment execution, grading, statistical analysis, tracing, optimization-lever plugins, and a locally hosted dashboard. Partner directly with enterprise clients to run staged campaigns, quantify cost-quality tradeoffs, project dollar impact, assess generalization, and deliver evidence-based recommendations. Provide technical leadership, establish evaluation standards, mentor engineers, and communicate findings to technical and business stakeholders.

Leads large-scale AI, digital, and cloud transformation programs from strategy through launch and realized business value. The role coordinates concurrent Agile teams, technical direction, architecture discussions, AI lifecycle governance, risk management, KPIs, stakeholder communication, staffing, and executive decision-making. It requires translating business goals into technical plans, managing complex dependencies and trade-offs, mentoring delivery leaders, and shaping delivery practices while partnering with clients on growth opportunities.

Train and operate machine learning models on Amazon SageMaker using AWS Trainium. Develop and optimize PyTorch training code for NeuronCore hardware, diagnose device and compiler issues, tune distributed workloads for throughput and cost, and deliver production training pipelines while collaborating directly with client and engineering teams.

Leads the architecture, development, optimization, and production deployment of complex AI and machine learning systems. Builds scalable ML platforms, pipelines, and infrastructure; establishes engineering standards; guides responsible AI practices; and mentors engineers. Partners with product and leadership teams to shape AI strategy, communicates technical tradeoffs, and drives high-stakes initiatives focused on reliability, scalability, security, cost, and safety. The role requires deep generative AI, agentic systems, RAG, cloud, and production engineering expertise.

Lead the architecture, development, deployment, and adoption of production AI/ML systems within client environments. Build scalable ML platforms, agentic systems, RAG pipelines, and integrations; resolve complex technical challenges; establish responsible AI and engineering standards; communicate tradeoffs to stakeholders; support client enablement; and mentor engineers. The role requires deep expertise in generative AI, AWS, cloud infrastructure, observability, security, cost optimization, Docker, and Kubernetes.

Leads the technical vision and delivery of an enterprise agentic AI program. Oversees architecture, evaluation, deployment, MLOps, reliability, security, and scalability across multiple workstreams. Partners with executives and client stakeholders on roadmaps and high-stakes decisions, aligns data science and engineering teams, establishes technical standards, and mentors senior technical leaders. Requires deep experience in AI/ML, LLMs, agentic systems, AWS, distributed systems, evaluation, and production-scale software engineering.

Train and operate production machine learning models on Amazon SageMaker using AWS Trainium. Optimize PyTorch code for Trainium compilation, memory, throughput, and distributed training; diagnose hardware, compiler, device, data, and code issues; build cost-aware end-to-end training pipelines; and collaborate directly with client and internal engineering teams to deliver production workloads.