Junior Forward-Deployed Software Engineer

You will work directly with Carma’s leadership and operating teams to build and improve the systems that run the business. The work will span full-stack application development, internal tools, AI workflow automation, data pipelines, AWS infrastructure, integrations, and rapid prototyping.

This is a strong fit for someone who can move quickly, learn unfamiliar domains, talk to non-technical users, and ship practical software that solves real problems. You do not need years of professional experience, but you should have meaningful project experience and the ability to figure things out independently.

The person we want is hungry, low-ego, and unusually capable for their stage. You may be a student, recent graduate, self-taught builder, hackathon person, research assistant, startup intern, or the kind of person who has been shipping side projects for years because you could not help yourself.

What You’ll Work On

• Build across Carma’s internal CRM, operating platform for submissions, applications, accounts, policies, agents, agencies, contacts, documents, audit logs, renewals, and post-bind workflows.
• Improve the quote-to-bind workflow: submission intake, underwriting status changes, brokerage routing, bind conversion, policy creation, post-bind checklist completion, renewal creation, and operational follow-up.
• Improve the broker intake experience, including form validation, field mapping, PDF/application generation, Cloudflare Pages, AWS Lambda ingest, SharePoint document handoff, and C360 submission creation.
• Extend AI-assisted workflows inside CRM: chat, Model Context Protocol (MCP) tools, stale-deal filters, pipeline summaries, task completion, document/activity context, suggestion prompts, and safe tool/action limits.
• Build operational dashboards and views across internal systems My Tasks, turnaround-time tracking, pipeline summaries, quote conversion, renewals, broker/agency activity, risk warnings, portfolio views, and executive reporting.
• Help maintain the platform foundation: Auth Service, Microsoft SSO, RBAC, API keys, shared cookies, audit trails, Microsoft Graph/Teams integrations, AdvanceHQ sync, and cross-service permission boundaries.
• Help maintain AWS and deployment infrastructure: Terraform, EC2, RDS, ALB, OpenVPN, Lambda, CloudWatch logs, backups, sandbox environments, deploy workflows, and prod-to-sandbox data seeding.
• Own the full user-visible surface of a change: backend data shape, frontend component, migration, permission boundary, tests, docs, and verification evidence.

Example Projects

• Build an end-to-end broker intake improvement: FirstConnect form change, Lambda payload mapping, C360 application/submission creation, SharePoint document upload, generated PDF update, audit log entry, and visible C360 UI confirmation.
• Add an underwriting or operations queue that surfaces stale submissions, overdue activity tasks, turnaround-time risk, missing documents, quote follow-ups, bind blockers, and post-bind checklist gaps.
• Improve the C360 AI assistant with new MCP tools such as “show stale deals,” “summarize my pipeline,” “list my tasks,” or “complete this activity,” then add evals and guardrails so the tools are safe to trust.
• Build a C3 risk or portfolio view that reads from C360, joins external risk data, shows account/submission pins on a map, flags coverage gaps, and explains why an account is critical, warning, or monitor.
• Add a dashboard tile or report that starts from a real operating question, verifies the source table or view, implements the backend query, renders the frontend component, and documents the metric semantics.
• Improve Auth/RBAC behavior across C360 and C3 so users only see the modules and records they should see, with backend enforcement and tests that prove hidden data is absent from response JSON.
• Build an Eng-Agent workflow that turns repeated review feedback into a checkable rule, eval fixture, retrieval packet, or pre-PR checklist item so the same class of mistake gets caught earlier next time.
• Harden deployment and observability for a real workflow: logs, health checks, cron behavior, sandbox flags, CloudWatch visibility, migration receipts, and a rollback or verification runbook., * You like being handed a real business problem, not a perfectly scoped ticket.
• You are comfortable talking to operators, underwriters, brokers, and leadership, then turning what you hear into software.
• You can move fast without being sloppy.
• You want broad exposure across product, engineering, AI, infrastructure, data, and insurance operations.
• You would rather ship a useful internal tool this week than polish a theoretical architecture for a month.
• You are excited by the idea that every project should leave the company smarter than before: better docs, better tests, better evals, better workflows.

Why This Is a Good Early-Career Role

This role offers unusual exposure for a junior engineer. You will not be sitting far away from the problems you are solving. You will see how an insurance MGA actually operates, where the manual work lives, and how software and AI can create leverage inside a real business.

You will get to build across the stack, work directly with decision-makers, and own projects that matter. The best person for this role is excited by a mix of engineering, product thinking, operations, AI, and infrastructure.

For the right person, this is closer to a founder-apprenticeship engineering role than a traditional entry-level job. You will learn how a company works by building the systems that make it work., * A short note on why this role interests you and what you have built that shows you can handle it.

• Links to projects, GitHub repos, demos, apps, research, or technical work you are proud of.
• Optional: a short screen recording or write-up of a project where you used AI, automation, infrastructure, or full-stack engineering to solve a real problem.

Ideal Candidate: Student, recent graduate, or early-career engineer with strong full-stack, AI, and infrastructure instincts, We are looking for a junior forward-deployed software engineer who wants to build close to the business. This is a hands-on engineering role for someone who likes turning messy operational workflows into clear tools, automations, dashboards, integrations, and AI-assisted systems., * Strong full-stack development ability, with experience building real applications beyond coursework.

• Comfort working across front end, back end, databases, APIs, authentication, deployment, and debugging.
• Practical AI experience, such as building with LLM APIs, retrieval workflows, structured extraction, tool calling, prompt evaluation, or AI-enabled internal tools.
• Working knowledge of AWS infrastructure, including services such as S3, Lambda, ECS, RDS, CloudWatch, IAM, API Gateway, or similar cloud building blocks.
• Ability to work with messy real-world data, business rules, documents, spreadsheets, and operational edge cases.
• Strong product instincts: you can ask good questions, understand the user, and build the simplest thing that actually works.
• Evidence discipline: before building, you ask where the data lives, which system owns the truth, who experiences the problem, and what would prove the fix worked.
• Shipping discipline: you test your work, show receipts, document what changed, and do not confuse a demo with production-ready software.
• Good written communication and comfort explaining technical tradeoffs to non-technical teammates.
• High ownership, curiosity, and willingness to work in a fast-moving startup environment.

Preferred Qualifications

• Experience with agent orchestration, multi-agent workflows, autonomous task execution, evals, tool routing, or workflow automation frameworks.
• Experience deploying AI applications in production or semi-production environments.
• Experience with document processing, OCR, PDF parsing, email automation, CRM workflows, or insurance/financial services operations.
• Comfort with TypeScript, React, Node.js, Python, SQL, Postgres, Docker, GitHub Actions, or modern serverless/cloud infrastructure.
• Prior startup experience, founder-led projects, open-source work, research projects, hackathon wins, or independently shipped products.