Applied AI/ML Engineer

7Seventy; United States; Hybrid; USD 175,000 - 250,000; Full time

Benefits offered by this job

Health insurance

Dental insurance

Vision insurance

Remote-first work environment

Conference travel budget

Flexible paid time off

Regular off-site events

Job summary

7Seventy is seeking an Applied AI/ML Engineer to own production AI products powered by large language models. You will drive the full lifecycle from prototype to production, focusing on performance, scalability, and reliability across GPU infrastructure.

The role emphasizes delivering production-ready AI systems, not research, collaborating across ML, inference optimization, GPU infra, and product teams to improve quality, throughput, and latency while managing end-to-end workflows.

Qualifications

• 3+ years of production AI/ML systems experience.
• Hands-on with LLM serving using vLLM or SGLang.
• Experience with RL or PT techniques (GRPO, PPO, DPO, SFT).

Responsibilities

• Own AI features from prototype to production deployment.
• Deploy and optimize LLM inference for throughput with low latency.

Skills

Independent worker

Problem solving

Collaboration

Tools

PyTorch

vLLM

SGLang

Megatron-LM

TensorRT-LLM

Kubernetes

Ray

Slurm

GitHub

Job description

Salary: $175,000 - $250,000 per year, plus equity

About the Role

This opportunity is for an Applied AI/ML Engineer responsible for building and delivering production AI products powered by large language models and modern machine learning technologies. You will own the complete development lifecycle, from prototype through production deployment, while optimizing model performance, inference efficiency, scalability, and operational reliability across GPU infrastructure.

This is a highly autonomous engineering role focused on delivering production-ready AI systems rather than research projects. You will work across machine learning, reinforcement learning, inference optimization, GPU infrastructure, and product development while continuously improving quality, throughput, latency, and operational efficiency.

What You'll Do

End-to-End AI Product Delivery

• Own AI-powered features and products from initial prototype through production deployment, including model selection, inference serving, evaluation, optimization, and continuous iteration.
• Deliver production-ready AI software designed for real-world performance rather than research prototypes.

Inference Serving

• Deploy and optimize large language model (LLM) inference using vLLM and SGLang.
• Optimize continuous batching, KV-cache management, quantization, speculative decoding, and multi-model routing to maximize throughput while minimizing latency and inference cost.
• Build, deploy, and maintain reinforcement learning and post-training pipelines using slime, Megatron-LM, SGLang, prime-rl, and related tooling.
• Design reward functions and verifier systems.
• Manage rollout orchestration, weight synchronization, and long-running training stability.
• Develop evaluation frameworks and benchmarking systems that measure model quality, throughput, and cost.
• Use evaluation data to continuously improve production AI systems.

Cross-Functional Engineering

• Partner with infrastructure engineers to improve GPU scheduling and fleet utilization.
• Collaborate with product teams to prioritize and deliver new AI capabilities.

Qualifications

• 3+ years of experience designing, deploying, and operating production AI or machine learning systems.
• Hands-on experience serving large language models using vLLM, SGLang, TensorRT-LLM, or comparable inference platforms.
• Experience with reinforcement learning or post-training techniques, including GRPO, PPO, DPO, SFT, or equivalent approaches.
• Strong PyTorch experience.
• Working knowledge of GPU execution, including batching, memory optimization, and CUDA fundamentals.
• Ability to work independently with minimal supervision while operating effectively in ambiguous, fast-moving environments.
• Experience with slime, prime-rl, the Verifiers library, or Megatron-LM is preferred.
• Experience with distributed training technologies, including FSDP and tensor, pipeline, or data parallelism.
• Experience with quantization, speculative decoding, or production inference optimization.
• Experience building verifiable inference systems or large-scale distributed platforms.
• Experience with Kubernetes and containerized deployments.
• Familiarity with GPU orchestration platforms such as Ray, SkyPilot, or Slurm.
• A public GitHub profile is required.
• Your GitHub profile must demonstrate at least one year of development activity.
• Applications without a qualifying public GitHub profile will not be considered.
• Health, dental, and vision insurance.
• Flexible paid time off.
• Conference travel budget.
• Remote-first work environment.
• Regular company off-site events.
• High-trust, high-performance engineering culture.