
Senior AI Software Engineer
AnaVation
- Chantilly, VA
- Permanent
- Full-time
- Clearance: Active TS/SCI within last 24 months
- Education: BA/BS is Computer Science or another related field
- Experience: BS + 10 Yrs or MS + 8 Yrs experience in computer science, AI, Machine Learning, or a related field.
- 5+ years of experience in AI/ML development, with at least 2 years focused on Agentic AI or autonomous systems.
- Proven track record of deploying production-grade AI systems, including framework experience such as AWS Bedrock.
- Strong problem-solving skills and ability to work in a fast-paced, collaborative environment.
- PhD in computer science, AI, Machine Learning, or a related field
- Technology Stack:
- Programming Languages: Python (primary), JavaScript/TypeScript (for API development), C++ (for performance-critical components).
- Frameworks and Libraries:
- Machine Learning: PyTorch, TensorFlow, JAX.
- Reinforcement Learning: Stable-Baselines3, Ray RLlib, OpenAI Gym, or Gymnasium.
- Agentic AI Frameworks: LangChain, LlamaIndex, AutoGen, or CrewAI.
- API Development: FastAPI, Flask, or
- Containerization: Docker, Kubernetes.
- Version Control: Git, GitHub, or GitLab.
- Databases: SQL (PostgreSQL, MySQL), NoSQL (MongoDB, DynamoDB).
- DevOps Tools: CI/CD pipelines (Jenkins, GitHub Actions), monitoring tools (Prometheus, Grafana).
- AI Models and Techniques:
- Large Language Models (LLMs): Experience with models like LLaMA, GPT, BERT, or Grok for natural language understanding and generation, including leveraging AWS Bedrock for LLM deployment and management.
- Reinforcement Learning (RL): Expertise in RL algorithms (e.g., DQN, PPO, SAC) and multi-agent RL systems.
- Agentic AI Paradigms:
- Knowledge of goal-driven agents, task decomposition, and autonomous planning (e.g., ReAct, Plan- and-Execute architectures).
- Prompt Engineering:
- Designing prompts for LLMs to achieve reliable and context-aware outputs, optimized for Bedrock's model ecosystem.
- Model Fine-Tuning:
- Techniques like LoRA, QLoRA, or full fine-tuning for domain-specific applications, with experience using Bedrock for fine-tuning workflows.
- Evaluation Metrics:
- Familiarity with BLEU, ROUGE, perplexity, and custom metrics for agent performance.