About the job
Join Our Mission at Twenty
At Twenty, we are committed to tackling one of the most pressing challenges of our era: safeguarding democracies in the digital landscape. We create innovative technologies that merge the realms of cyber and electromagnetic domains, where the pace of operations surpasses human perception and complexity goes beyond traditional limits. Our team not only addresses problems but also delivers transformative outcomes that significantly enhance national security. We embody a spirit of pragmatic optimism, believing that while our mission to protect America and its allies is formidable, it is indeed achievable.
Position Overview
As an Applied AI Engineer, you will develop and implement language-model-driven systems that bolster Twenty's vital cyber capabilities for U. S. national security. You will manage the complete workflow, from curating specialized datasets and post-training models to deploying dependable inference and retrieval systems in live environments. Collaborating closely with product and engineering teams, you will convert real operational demands into high-performance AI features, operating seamlessly across cloud and on-premises environments where speed, accuracy, and security are paramount.
Your Profile
You are driven by tangible outcomes and aspire for your work to directly influence national security initiatives.
You prioritize rigor: clean data, quantifiable evaluations, and reproducible experiments take precedence over mere demonstrations.
You strike a balance between research curiosity and product intuition, you ship, analyze, iterate, and refine.
You are adept at navigating both cloud and on-premises constraints and can adjust to various environments.
You convey ideas clearly to both technical and non-ML stakeholders, and your documentation is user-friendly.
You understand systems thinking: models, retrieval, infrastructure, and feedback loops must function cohesively.
You excel in dynamic teams with high standards, open feedback, and a strong sense of ownership.
Key Responsibilities
Develop, clean, and sustain high-quality training and evaluation datasets tailored for specialized AI applications.
Fine-tune language models (ranging from small specialized to medium foundation models) according to mission requirements.
Apply post-training and alignment techniques to enhance task performance and reliability.
Construct retrieval-augmented generation (RAG) systems that integrate model outputs with external knowledge.
Enhance and optimize model serving infrastructure for production deployment.
