Software Development Engineer – Agentic AI

Trellix•Frisco, TX

About The Position

Trellix is hiring a Software Development Engineer to lead the design, development, and operation of a generative AI platform that powers AI capabilities across its cybersecurity product portfolio. The role builds production agentic systems, including multi-agent orchestration and autonomous SOC workflows, with deployments in federal environments. The engineer owns platform features from design and development through testing, automation, and ongoing production operations. The position focuses on building secure, resilient, and performant AI solutions; developing and hardening agent workflows; and delivering work to project goals and customer commitments. It also involves implementing OpenTelemetry instrumentation, contributing to operational dashboards, documenting technical designs and architectural decisions, and owning production incidents through triage, root-cause analysis, post-mortems, and SLA commitments. The role requires collaboration across engineering, product, and security research teams, along with clear communication of technical decisions to stakeholders.

Requirements

At least 5 years of professional experience in Python application software development.
Demonstrated experience building and operating production AI or platform systems.
Linux proficiency.
Excellent Python development and debugging skills.
Strong understanding of data structures and design patterns.
Proficiency with REST and asynchronous Web APIs.
Experience with CI/CD pipelines, such as GitHub Actions or equivalent.
Strong written communication skills for design documentation and asynchronous collaboration.
Ability to work autonomously in a fast-moving, ambiguous environment.
Hands-on experience with large language models in production.
Experience with LangChain, including building and operating stateful multi-agent workflows.
Experience with prompt orchestration and chain composition.
Familiarity with agentic AI concepts and patterns, including ReACT, chain-of-thought, tool use, and Deep Agents.
Experience deploying and operating vLLM for self-hosted inference.
Familiarity with MCP (Model Context Protocol) for agentic tool integration.
Excellent communication and collaboration skills.

Nice To Haves

Shell scripting.
FastAPI.
Node.js and TypeScript for tooling and API integration layers.
Postgres.
Knowledge graphs, including NebulaGraph or equivalent.
Vector databases, including Qdrant or equivalent.
Embedding pipeline experience, including chunking strategies and retrieval tuning.
Redis.
Langfuse or LangSmith for agent observability.
Kubernetes.
AWS services including RDS, EKS, ElastiCache, and Bedrock.
Working knowledge of threat detection, EDR telemetry, SOC workflows, or SIEM platforms.
Understanding of Security Incident and Event Management (SIEM) and incident response.

Responsibilities

Lead design and development of the generative AI platform, including core functionality, agentic workflows, and platform-level features from concept through production.
Own features and functions from initial design and development through testing, automation, and ongoing operational health.
Build, iterate on, and harden multi-agent pipelines, including tool use, inter-agent coordination, and autonomous decision workflows for security operations.
Deliver solutions on time, within budget, and to required quality standards, meeting project goals and customer commitments.
Implement best practices for application resilience, security, and performance, with particular attention to sensitive security operations data.
Design and implement OpenTelemetry instrumentation, contribute to operational dashboards, and provide platform health and usage insights to engineering leadership and stakeholders.
Analyze feature requirements and prepare design documentation, architectural decision records, and async-friendly technical specifications.
Own production issues end-to-end, including triage, root-cause analysis, post-mortems, and SLA commitments.
Collaborate across engineering, product, and security research teams and communicate technical decisions and tradeoffs to non-engineering stakeholders.