Software Engineer – Applied AI

Overview

In this role you will help accelerate research and production workflows by building scalable AI infrastructure and agent systems at Hudson River Trading. You will work across research engineering, post-training methods, and deployment to deliver reliable, safe AI capabilities that scale with the firm’s complex research and trading workloads. You’ll collaborate with cross-functional teams to turn experimental results into durable systems and embed rigorous evaluation in decision making. This is an opportunity to shape autonomous AI that supports researchers and developers at scale, with meaningful impact on the trading floor.

Pay / Benefits

• discretionary performance-based bonuses
• competitive benefits package

Responsibilities

• Build scalable infrastructure for training, evaluating, and operating language models and AI agents
• Develop research environments, evaluation systems, and experimentation platforms to measure and improve agent performance
• Improve model and agent capabilities via post-training, reinforcement learning, prompting, tool design, and workflow optimization
• Design sandboxing, policy enforcement, monitoring, and safeguards for autonomous systems
• Partner with cross-functional teams to identify new AI-enabled work and turn experiments into durable systems
• Evaluate new models, techniques, and products and inform buy/build/adapt decisions
• Share knowledge and establish practices for building with AI across the company

Key requirements

• Bachelor’s degree in Computer Science, Engineering, or a related field, or equivalent practical experience
• Strong Python proficiency
• Excellent software design, debugging, and problem-solving skills
• Experience building and operating substantial software systems
• Demonstrated engagement with modern AI systems, tools, or research
• Machine learning research or engineering including post-training, fine-tuning, reinforcement learning, or alignment
• AI-agent architecture, evaluation, or tool-use environments
• Distributed systems, model serving, orchestration, or high-throughput research infrastructure
• Security sandboxing, policy enforcement, adversarial testing, or governance for AI systems
• Forward-deployed engineering: working closely with users to discover, build, and operationalize new capabilities
• Linux, systems performance, networking, C++, or TypeScript

• clear communication
• curiosity about capabilities, limitations, and risks of autonomous systems
• quick adaptation across domains
• Python
• C++
• TypeScript

…