Applied AI Engineer

About the Company

America is under sustained cyber attack. Our adversaries infiltrate our networks, steal our IP, and degrade the digital infrastructure that modern life runs on. They’ve learned—correctly—that those attacks rarely produce consequences.

Twenty was founded to change that, by making our adversaries think twice before they attack us. Our vision is American and allied primacy in cyberspace—a future where they cannot contest us, deterrence is assured, and the free world remains secure.

Founded in 2024, Twenty Technologies ((Use the "Apply for this Job" box below).) industrializes offensive cyber operations for the U.S. and its allies. Headquartered in Arlington, Virginia, Twenty has raised $38M from Caffeinated Capital, General Catalyst, and In‑Q‑Tel.

You’ll build and ship language-model-powered systems that strengthen Twenty’s mission‑critical cyber capabilities for U.S. national security. You’ll own the end‑to‑end workflow—from curating specialized datasets and post‑training models to deploying reliable inference and retrieval systems in production. You’ll partner closely with product and engineering to translate real operational needs into high‑performing AI features, operating across cloud and on‑premises environments where speed, correctness, and security matter.

• You’re motivated by real‑world outcomes and want your work to directly impact national security missions.

• You care about rigor: clean data, measurable evaluation, and repeatable experiments beat demos.

• You balance research curiosity with product instincts—you ship, observe, iterate, and harden.

• You’re comfortable working across cloud and on‑premises constraints and adapting to the environment.

• You communicate clearly with engineers and non‑ML partners, and you write documentation people use.

• You think in systems: models, retrieval, infrastructure, and feedback loops all have to work together.

• You thrive in fast‑moving teams with high standards, direct feedback, and high ownership.

• Create, clean, and maintain high‑quality training and evaluation datasets for specialized AI use cases.

• Fine‑tune language models (small specialized through medium foundation models) for mission needs.

• Implement post‑training and alignment approaches to improve task performance and reliability.

• Build retrieval‑augmented generation (RAG) systems that ground model outputs in external knowledge.

• Develop and optimize model serving infrastructure for production deployments.

• Design evaluation frameworks and test harnesses to measure quality, latency, and regressions.

• Integrate AI capabilities into applications and workflows using modern orchestration frameworks.

• Collaborate with cross‑functional partners to identify high‑leverage use cases and deliver solutions.

• Produce clear technical documentation for models, datasets, and operational processes.

• You have 4+ years of professional software development experience building and supporting ML/AI‑enabled applications.

• You have strong Python skills and deep learning experience with PyTorch, Tensor Flow, or JAX.

• You have hands‑on experience with LLM post‑training methods (e.g., continued pre‑training, SFT, RLHF, DPO, PPO, GRPO).

• You have experience curating, cleaning, and preprocessing datasets for training and evaluation.

• You have working knowledge of relational, graph, and vector database concepts.

• You have experience designing or using evaluation metrics and testing procedures for LLMs and agents.

• You have experience integrating LLM/agent systems using frameworks like Pydantic‑AI, Lang Chain/Lang Graph, or CrewAI.

• You have a Bachelor’s degree in Computer Science, Software Engineering, or a related field (or equivalent practical experience).

• You have deployed models to production and supported them through real‑world usage and incidents.

• You have experience with distributed training systems and performance debugging at scale.

• You have implemented quantization or other optimization techniques to improve inference efficiency.

• You have strong prompt engineering and model alignment instincts for reliability and control.

• You have experience building MLOps/LLMOps/Agent Ops practices (versioning, rollout,…