Lead Application Developer - Agentic AI (Relocation candidates are welcome)

Candidates that are willing to relocate to Tampa can apply

we have 2 roles:

One is a Java with Agentic AI

One is Python with Agentic AI

Skills
- Proficiency in Java or Python (with experience in TypeScript, Go a plus).
- Solid understanding of AWS architecture and services—deployment, monitoring, security, and cost
optimization.
- Familiarity with agentic or LLM frameworks such as Strands, CrewAI, LangGraph, Agent Core, or
similar.
- Experience with retrieval systems, vector databases, embeddings, and orchestration frameworks.
- Strong grounding in secure API design, data modeling, and CI/CD automation.
- Proven record of writing clean, testable, production-grade code.
- Experience designing tool/function-calling integrations and Model Context Protocol (MCP) servers
and connectors.
- Strong prompt and context engineering, including agent memory, state, and session/context
management.
- Experience with agent and LLM evaluation, guardrails, and safety controls—hallucination
mitigation, content filtering, and human-in-the-loop oversight.
- Hands-on experience with foundation models via Amazon Bedrock (e.g., Anthropic Claude),
including model selection and prompt/parameter tuning for latency and cost.
- Familiarity with multi-agent orchestration and workflow design—planning, task decomposition,
routing, tool use, and error handling.
- Experience with containerization and orchestration (Docker, Kubernetes) for deploying scalable,
resilient services.
- Knowledge of Terraform for IaC deployments. Hands-on experience with Terraform is a plus.
- OpenTelemetry for Dynatrace and Observability
- Experience with agent performance tuning and cost optimization.
- Experience with time-boxed A2A and Swarm/Crew Agentic solutions
- Deploying large language models on-prem
- Experience with fine-tuning LLMs - nice to have
- Mindset: Deep curiosity, bias toward execution, and respect for precision and reliability.
- Financial services experience preferred but not required.