Position Summary:We're looking for a Junior Software Engineer to help take various Agentic verification capabilities from proof-of-concept to a real, production system. This is a great fit for a recent or soon-to-graduate Master's student who has spent their program building real projects with modern AI tools - not just using ChatGPT, but experimenting with coding assistants and building agent-style projects on your own initiative. We're not expecting years of professional AI experience. We're looking for people who are curious, self-driven, and have already started exploring this space through coursework, personal projects, or hackathons, and who want to keep building in a production environment.
This role is expected to leverage AI and emerging technologies to improve productivity, enhance decision-making, and continuously optimize how work is performed.
This will be a full-time, permanent position working out of the San Diego office on a hybrid model (3-days per week in the office).
What You'll Be Doing:- Help take various Discovery Agent prototypes (e.g. Vertex AI-based) through the SDLC - contributing to design, implementation, testing, and deployment (via API and Agentic access).
- Help harden POC-quality agent code toward production quality: error handling, observability, testing, and basic cost/latency awareness.
- Help build test cases and evaluation checks for agent behavior - an important and different skill from testing traditional deterministic code.
- Support data ingestion and pipeline work feeding the agent, using AWS/BigData tooling.
- Work with Product team and senior engineers to turn business requirements into shipped features.
- Learn and contribute to production support practices: monitoring, alerting, and incident response for an AI-driven service.
What You'll Bring:- 1-2 years of experience working in a software engineering or development environment.
- Master's degree in Computer Science, Data Science, Engineering, or a closely related field.
- Solid foundation in Python and/or another backend language, with the ability to write clean, readable, testable code.
- Understanding of SQL and/or NoSQL databases, and basic API design (REST at minimum).
- Exposure to cloud infrastructure (AWS preferred) - coursework, internship, or personal project experience is fine.
- Some exposure to data pipelines or data processing (e.g., via coursework, internships, or projects using tools like PySpark, Airflow, or similar) is a plus, not a requirement.
- A self-starter mindset - comfortable with ambiguity, asking good questions, and figuring things out rather than waiting for a fully specified task.
- Genuinely excited about working with the newest AI tooling - you've already been experimenting with it because you wanted to, not because a class required it.
- Driven to actually ship usable products: you'd rather get a real project working end-to-end than perfect just one piece of it.
- Eager to learn how production systems really work - testing, deployment, monitoring, on-call support and troubleshooting- beyond what a class or internship typically covers.
- A good collaborator who communicates clearly and is comfortable working alongside more senior engineers. Bonus (Nice to Have, Not Required)
- Coursework or projects touching orchestration tools (Apache Airflow, Dagster), OpenSearch/Elasticsearch, or BigData tools (Spark, Presto/Trino, Hive).
- Familiarity with TypeScript/React/Node.js for full-stack project work.
- Exposure to infrastructure-as-code (Terraform, AWS CDK).
- Interest or coursework in identity verification, fraud, security, or fintech.
Level-Setting on AI Experience:- You don't need professional experience shipping AI products - we're looking for genuine curiosity and self-directed learning and hands-on experimentation, not a resume full of AI job titles.
- Comfortable using AI coding assistants (e.g., Claude Code, GitHub Copilot, Cursor) as part of how you write code, and interested in getting better at it.
- Have built at least one project - for coursework, a hackathon, research, or on your own - involving an LLM doing more than single-shot Q&A: calling tools/functions, chaining steps, using retrieval, or coordinating multiple steps toward a goal.
- Exposure to LLM APIs (Anthropic, OpenAI, Google Vertex AI/Gemini) or agent frameworks (LangGraph, LangChain, AutoGen, or similar) is a strong plus, even at a project or coursework level - deep production expertise is not expected.
- Genuinely interested in the practical side of AI engineering: why these systems behave unpredictably, and what it takes to make them reliable enough to ship.
