Asks for

The ad says“You have experience integrating LLM/agent systems using frameworks like Pydantic-AI, LangChain/LangGraph, or CrewAI.”
The ad says“You have strong Python skills and deep learning experience with PyTorch, TensorFlow, or JAX.”
The ad says“You have hands-on experience with LLM post-training methods (e.g., continued pre-training, SFT, RLHF, DPO, PPO, GRPO).”
The ad says“You have working knowledge of relational, graph, and vector database concepts.”
The ad says“You have hands-on experience with LLM post-training methods (e.g., continued pre-training, SFT, RLHF, DPO, PPO, GRPO).”
The ad says“You have strong Python skills and deep learning experience with PyTorch, TensorFlow, or JAX.”
The ad says“You have experience integrating LLM/agent systems using frameworks like Pydantic-AI, LangChain/LangGraph, or CrewAI.”
The ad says“You have experience integrating LLM/agent systems using frameworks like Pydantic-AI, LangChain/LangGraph, or CrewAI.”
The ad says“You have experience designing or using evaluation metrics and testing procedures for LLMs and agents.”
The ad says“You have hands-on experience with LLM post-training methods (e.g., continued pre-training, SFT, RLHF, DPO, PPO, GRPO).”
The ad says“You have hands-on experience with LLM post-training methods (e.g., continued pre-training, SFT, RLHF, DPO, PPO, GRPO).”
The ad says“You have experience integrating LLM/agent systems using frameworks like Pydantic-AI, LangChain/LangGraph, or CrewAI.”
The ad says“You have strong Python skills and deep learning experience with PyTorch, TensorFlow, or JAX.”
The ad says“You have strong Python skills and deep learning experience with PyTorch, TensorFlow, or JAX.”
The ad says“You have working knowledge of relational, graph, and vector database concepts.”
The ad says“You have hands-on experience with LLM post-training methods (e.g., continued pre-training, SFT, RLHF, DPO, PPO, GRPO).”
The ad says“You have hands-on experience with LLM post-training methods (e.g., continued pre-training, SFT, RLHF, DPO, PPO, GRPO).”
The ad says“You have strong Python skills and deep learning experience with PyTorch, TensorFlow, or JAX.”
The ad says“You have working knowledge of relational, graph, and vector database concepts.”

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 (www.twenty.io) industrializes offensive cyber operations for the U.S. and its allies. Headquartered in Arlington, Virginia, Twenty has raised $168M from Khosla Ventures, Accel, Caffeinated Capital, Friends & Family Capital, Point72 Ventures, General Catalyst, and In-Q-Tel.

Mission | On Site | Full Time | U.S. Citizenship Required / No Active Clearance Required

Role Summary

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.

What You’ll Do

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.

Who You Are

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.

Must Have

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, TensorFlow, 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, LangChain/LangGraph, or CrewAI.
You have a Bachelor’s degree in Computer Science, Software Engineering, or a related field (or equivalent practical experience).

Nice to Have

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/AgentOps practices (versioning, rollout, monitoring).

Tech Environment (You Might Work With)

Deep learning stacks: PyTorch, TensorFlow, JAX
LLMOps and serving: vLLM, TensorRT, ONNX
Retrieval and storage: pgvector, ChromaDB, Pinecone, Milvus, Weaviate; relational/graph databases
Orchestration: Pydantic-AI, LangChain/LangGraph, CrewAI
Infra: Docker, Kubernetes; cloud platforms (AWS, GCP, Azure)
Experiment and artifact tracking: dataset/prompt/model versioning

Security / Work Environment

Must be eligible to obtain and maintain a U.S. Government security clearance.

Benefits
What's on the table:

Health. Medical, dental, and vision plan options. Life / AD&D, disability coverage options.
Family. Paid parental leave for eligible full-time employees. 12 weeks for birthing parents, 4 for non-birthing parents, 6 weeks for adoptive, foster, or intended parents through surrogacy.
Vacation. Paid holidays and flexible PTO. Take what you need.
Retirement. 401(k) with pre-tax and Roth options. HSA/FSA options, dependent care FSA.

Benefits vary by location, role, and eligibility. Full plan details provided during the interview and offer process.

If this role sounds like you, apply and share with us your interest

Due to U.S. government contract and security requirements, this role is limited to U.S. citizens. Some positions may also require eligibility to obtain and maintain a U.S. Government security clearance. Any active clearance requirement will be listed in the role description.

Twenty is an equal opportunity employer. We consider all qualified applicants without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, veteran status, disability, or any other protected status, consistent with applicable law.

If you need a reasonable accommodation during the hiring process, let us know and we will work with you.