Applied AI Researcher, Post-Training

Distyl Inc.

Job Overview

Location

Indiana, USA

Job Type

Full-time

Full Job Description

📋 Description

• Distyl AI is at the forefront of revolutionizing enterprise operations for Fortune 500 companies by developing and deploying production-grade AI systems. Our unique approach, bolstered by a strategic partnership with OpenAI, proprietary software accelerators, and deep-seated enterprise AI expertise, enables us to deliver tangible value and working AI solutions within a remarkably short timeframe, often within a single quarter. We empower businesses across diverse sectors, including insurance, consumer packaged goods, and non-profits, to identify, build, and fully realize the potential of their Generative AI investments, frequently marking their initial foray into this transformative technology.
• As a member of our dynamic team, you will play a pivotal role in guiding these enterprises through their AI journey. This involves not only understanding their core challenges but also architecting and implementing AI solutions that drive significant financial impact and demonstrably improve the daily experiences of their end-users. Our methodology is deeply rooted in a customer-centric philosophy, where we meticulously work backward from the customer's most pressing problems to ensure our solutions are not only innovative but also highly relevant and impactful.
• The Post-Training team, in particular, is dedicated to the critical task of adapting cutting-edge foundation models to meet the rigorous performance, safety, and alignment requirements of real-world enterprise applications. Researchers on this team are instrumental in developing and rigorously evaluating advanced techniques. This includes, but is not limited to, supervised fine-tuning, sophisticated preference optimization methods such as Direct Preference Optimization (DPO), Reinforcement Learning from Human Feedback (RLHF), and Reinforcement Learning from AI Feedback (RLAIF). Furthermore, the team explores and implements continual adaptation strategies to ensure models remain aligned with Distyl's complex enterprise systems and evolving business needs.
• The core mission of our Post-Training researchers is to bridge the gap between the raw, inherent capabilities of foundation models and the trustworthy, contextually relevant, and system-aligned behavior demanded by enterprise environments. This involves a deep dive into the intricate trade-offs inherent in AI development, such as balancing broad generalization with specialized task performance, optimizing data efficiency while maintaining robustness, and ensuring a delicate equilibrium between model capability and controllability.
• Your work will directly influence how Distyl leverages foundation models safely, effectively, and at an unprecedented scale across a multitude of industries. You will be tasked with investigating novel methodologies for aligning large language models (LLMs) with both human-defined objectives and complex system-level requirements. This research is crucial for building reliable and predictable AI systems that Fortune 500 companies can depend on for their most critical operations.
• We are seeking creative and forward-thinking researchers who are not content with merely achieving incremental improvements on existing benchmarks or optimizing established processes. Instead, we are looking for individuals who possess the vision and drive to fundamentally redefine how software is utilized through AI. While our researchers hail from diverse academic backgrounds, they share a common thread: a strong track record of impactful research, an innate ability to operate in an AI-native manner, and a desire to break free from the constraints of traditional research organizations.
• You will be instrumental in building intelligent systems that leverage AI models, rather than solely focusing on the training or fine-tuning of the models themselves. This requires a deep understanding of compound AI systems, the dynamics of agentic collaboration, and the application of associated techniques like ensembling, ReAct patterns, and Graph-of-Thoughts reasoning. Your ability to translate complex research concepts into practical, deployable solutions will be paramount.
• The ideal candidate will possess a proven track record of research achievements, demonstrated through publications in prestigious journals, impactful contributions shared on platforms like Twitter, or other significant research outputs. You should be someone who actively uses AI tools in your daily workflow, such as ChatGPT, Cursor, and Perplexity, to enhance your productivity and explore new possibilities. This hands-on approach is essential for understanding how AI can truly revolutionize workflows, both your own and those of our clients.
• Strong programming and data analysis skills are essential. While you may not identify strictly as a software engineer, you must possess the capability to prototype your innovative ideas and conduct rigorous experiments to validate their effectiveness, presenting compelling evidence to senior AI leaders within Fortune 500 organizations. We value a bias towards action and tangible results – our customers are eager to witness the immediate power of AI, not just discuss theoretical possibilities.

🎯 Requirements

• Deep understanding of post-training techniques for large language models, including supervised fine-tuning, preference optimization (RLHF, DPO), LoRA, PEFT, and instruction-tuning pipelines.
• Proven experience in adapting frontier foundation models (LLMs/SLMs) to specialized domains or specific behaviors through data curation, reward modeling, or continual pre-training.
• Experience building and deploying intelligent systems using AI models, focusing on compound AI systems, agentic collaboration, and techniques like ensembling, ReAct, or Graph-of-Thoughts, rather than solely on model training.
• A strong, demonstrable track record of research results, evidenced by publications in top-tier journals, significant contributions to open-source projects, or a compelling portfolio of innovative work.
• Daily, hands-on usage of AI tools such as ChatGPT, Cursor, Perplexity, or similar technologies to accelerate personal and professional workflows.
• Proficient programming and data analysis skills, with the ability to prototype ideas, design and execute experiments, and present findings effectively to technical and non-technical stakeholders.

🏖️ Benefits

• Competitive base salary range of $130K - $250K, commensurate with experience, location, and level.
• Meaningful equity in a high-growth, venture-backed startup.
• Comprehensive benefits package including 100% covered medical, dental, and vision insurance for employees and their dependents.
• 401(k) plan with additional perks, such as commuter benefits and in-office lunch provisions.
• Access to state-of-the-art AI models and generous usage of modern AI tools.
• Opportunity to work on high-impact projects with leading Fortune 500 enterprises.
• A mission-driven, fast-paced culture that values curiosity, pragmatism, and excellence.

Skills & Technologies

React

Rails

GitHub

Remote

Ready to Apply?

Apply Externally

You will be redirected to an external site to apply.

Distyl Inc.

Visit Website

About Distyl Inc.

Distyl is a cloud-native platform designed to simplify and accelerate the development and deployment of machine learning (ML) models. It provides a unified environment for data preparation, model training, versioning, and deployment, enabling data scientists and ML engineers to move from experimentation to production faster. The platform offers features such as automated data pipelines, managed training infrastructure, and scalable model serving. Distyl aims to reduce the complexity and operational overhead associated with MLOps, allowing organizations to focus on building and deploying impactful ML solutions. It supports various ML frameworks and integrates with existing cloud infrastructure.

View Company Profile