Security Engineering Intern - Summer 2026

Lambda•San Francisco, CA

2d•Hybrid

About The Position

Lambda Security protects some of the world's most valuable digital assets: invaluable training data, model weights representing immense computational investments, and the sensitive inputs required to leverage best of breed AI models. We're responsible for securing every byte that powers breakthrough artificial intelligence. As a Software Engineering Intern on the Security team, you'll spend the summer shipping production security detections against Lambda's real log data: code that fires on real events and routes to a real on-call team. This is not a sandbox project. The detections you build will be the first line of warning when something looks wrong in our environment. You'll work alongside a senior engineer mentor on code review, security concept training, and unblocking your work. Between those sessions, your primary partner is a frontier LLM: you'll use it to learn log schemas, reason through threat models, and accelerate your way through code you've never written before. We expect interns to drive their own learning and use AI tools fluently; the bar for what a great junior engineer can ship has changed, and this role is designed around that change. The work spans LLM-based classifiers for sensitive data exposure, identity and access pattern detection, and the security infrastructure that everything else depends on. You and your mentor will pick the specific projects together, prioritizing where the data is richest and the business value is clearest. If you love writing code, are excited about working with AI tools to solve problems you couldn't solve alone, and find the idea of protecting AI infrastructure genuinely interesting - even if you've never taken a security class - we'd love to hear from you. We value diverse backgrounds, experiences, and skills, and we are excited to hear from candidates who can bring unique perspectives to our team. If you do not exactly meet this description but believe you may be a good fit, please still apply and help us understand your readiness for this role.

Requirements

Currently enrolled in a Bachelor's or Master's program in Computer Science or a related field, returning to school after the summer.
Independent technical projects you've shipped: side projects, open source contributions, hackathons, research code, anything where you scoped a problem and built a working solution. Resumes that only list classwork are unlikely to stand out.
Fluency writing code in at least one general-purpose language (Go or Python preferred, but we care more about your ability to learn and leverage an LLM than your starting stack).
Comfort using frontier LLMs as a coding partner. You can tell when the model is wrong, push back on bad suggestions, and use it to accelerate learning rather than replace it.
Ability to take a fuzzy problem ("we need to know when someone overshares a sensitive document") and turn it into a scoped, shippable piece of work without needing a ticket pre-written for you.
Curiosity about how systems get attacked and defended, even if you've never formally studied security. We can teach you security; we cannot teach curiosity.
Strong written communication. Much of the work is reasoning in writing, asking sharp questions, and leaving behind code and documentation that the next person can use.

Nice To Haves

Prior software engineering internship.
Familiarity with cloud infrastructure (AWS, GCP, or similar), command-line environments, or systems-level programming.
You’ve built something with AI assistance that pushed beyond what you could have done alone.
Experience working with log data, structured event data, or building anything that processes streams of records.
Prior exposure to security concepts (a CTF you played, a security class you took, a side project you built, a security mistake you made and learned from, etc.).
Coursework or projects involving machine learning, data classification, or applied AI.

Responsibilities

Ship real production detections you develop against Lambda's real log data over the course of the summer, chosen with your mentor from a menu of high-value security gaps.
Use frontier LLMs as your primary partner for learning unfamiliar codebases, reading log schemas, reasoning through threat models, and accelerating your output as a working software engineer.
Build LLM-powered classifiers and rule-based detections that turn raw log data into actionable alerts for the Security team.
Work closely with your mentor in code reviews and threat hunting, and self-direct between them, using AI tools to investigate and sharpen your questions before bringing them to a human.
Document what you build clearly enough that the on-call engineer who responds to your alert at 2am understands what it means and what to do.
Present your work at the end of the summer to the security team and company leadership, walking through what you shipped, what you learned, and what you'd build next.