Forward Deployed Engineer II - Custom Model and Optimization

About the role

As a Forward Deployed Engineer II (FDE II) on the Custom Model and Optimization team, you embed directly with strategic customers to build, optimize and run custom generative AI models in production alongside their engineering teams. You own complete model workstreams end-to-end, from defining the evaluation bar and preparing data, through training and inference optimization, to production deployment. The team partners with enterprises, AI-native startups and foundation model providers, and the role requires up to 25% travel.

Key responsibilities

• Embed within customer ML and engineering teams, translating ambiguous goals into a measurable evaluation bar and technical plan covering model selection, customization approach and serving architecture.
• Customize and post-train open-weight and foundation models through continued pre-training, supervised fine-tuning, preference optimization and reinforcement learning (including RLHF, RL with verifiable rewards and multi-turn agentic RL), and build data pipelines, synthetic data generation and evaluation harnesses.
• Design and implement distributed training and RL post-training pipelines using frameworks such as PyTorch FSDP2, Megatron-LM, Miles and verl, and build RL interaction harnesses with NeMo Gym or OpenEnv.
• Develop production inference stacks with engines such as SGLang and vLLM, optimizing throughput, latency, memory footprint and cost per token through quantization, parallelism, speculative decoding and batching.
• Migrate and tune training and inference workloads for AWS Trainium using the AWS Neuron SDK, profiling performance and closing accuracy-parity gaps against GPU baselines.
• Codify reusable fine-tuning recipes, evaluation frameworks and benchmarking tools, and feed field learnings back to AWS service teams such as Amazon Bedrock, Amazon SageMaker AI and AWS Neuron.

Required profile

• 3+ years of non-internship professional software development experience.
• 2+ years in an Agile, startup or rapidly scaling environment.
• Hands-on experience with generative AI technology such as LLMs, vision-language models or speech models.
• Experience communicating technical concepts to technical and non-technical audiences.
• Preferred: Bachelor's degree in computer science or equivalent.

Required skills

• Programming in a modern language such as Java, C++ or C#, including object-oriented design.
• AI tools: eval frameworks, agent tooling, RAG pipelines and prompt engineering.
• Preferred: full software development lifecycle, including code reviews, source control, build processes and testing.
• Preferred: large-scale distributed LLM training on multi-node clusters (PyTorch FSDP2, Megatron-LM, NeMo, DeepSpeed, NeuronX Distributed).
• Preferred: LLM post-training (SFT, DPO, RLHF, RL with verifiable rewards or agentic RL) and inference optimization on AWS Trainium, Inferentia, SageMaker AI, SageMaker HyperPod or EKS, or custom kernels (CUDA, Triton, NKI).
• Preferred: deploying and operating ML models in customer production environments.