D
AI Engineer (Associate to Mid level)
DiversyFund, Inc
2 hours ago
Full-time
On-site
San Diego, California, United States
DiversyFund is hiring an associate to mid-level AI engineer to build the AI products that reach our investors and prospects: AI-driven email campaigns, AI texting, social media campaigns and the agents behind them. You'll work directly with our VP of Engineering and AI Product Development and our CTO/Chief Architect.
Our Chief Architect proves what's possible with AI. Your job is to turn those prototypes into reliable products that run every day, measure their results, and stay inside the rules that govern financial communications. This role suits someone early in their career who wants to prove themselves, or an experienced engineer who wants to start fast. You'll get hands-on mentoring, real ownership from week one, and a clear path to grow. Term Detail Type 13-week contract, with the opportunity to extend or convert to full time Rate $30 to $40 an hour, based on experience Hours Full time, 40 hours a week Start As soon as possible Location San Diego, CA, on-site Reports to VP of Engineering and AI Product Development What you'll do Build AI campaign products. Ship AI-written email, SMS and social media campaigns: audience selection, personalized drafts, scheduling, sending and reply handling. Build agents with guardrails. Design LLM agents with tool calling, structured outputs and clear limits on what they can do. Every bulk send, data change or external action goes through a human approval step. Build compliance in, not on. Respect consent, opt-outs, quiet hours and suppression lists for every message. Keep campaign content within the rules for financial communications. Measure quality with evals. Build evaluation sets for each agent, test prompt and model changes against them before release, and track accuracy, tone and policy violations. Connect to our platforms. Integrate with our outbound sending system, our CRM, email and SMS providers, and social media APIs. Track results. Report deliverability, replies, conversions and opt-outs, and use them to improve campaigns. Run AI in production. Log agent runs, monitor cost and latency, handle model failures and fallbacks, and keep secrets and personal data out of prompts and logs. Prototype new ideas. Turn the team's new AI product ideas into working, reviewed prototypes quickly. Review agent-written code. Much of our code is written with AI coding tools. You'll review and test what they produce in your area before it ships. What you need ~1 to 4 years of software engineering experience, including internships, contract work or substantial projects. ~ LLM applications: you've shipped something real on Claude, OpenAI or similar models: prompt design, tool calling, structured outputs and handling model failures. ~ Agentic systems: experience with agent frameworks or patterns such as the Vercel AI SDK, LangGraph, the OpenAI Agents SDK or MCP servers, and an understanding of where agents go wrong. ~ Evals: you know how to build test sets for model output and measure whether a change made it better or worse. ~ Programming: TypeScript (our main language) and some Python; Node.js and React or Next.js. ~ APIs and integrations: REST APIs, webhooks, OAuth and background jobs that retry safely. ~ Data: solid SQL with PostgreSQL; comfortable reading and shaping data for prompts and retrieval. ~ Cloud: working knowledge of AWS or Vercel and serverless deployment. ~ Git and code review: pull requests, CI and tests for the code you write. ~ AI-assisted development: you use tools like Claude Code or Cursor daily, and you catch their mistakes. ~ Judgment: you treat customer data and outbound messages with care, and you know when an agent should stop and ask a person.
Nice to have Fintech: investor communications, payments, KYC, or work under SEC or FINRA rules. Email or SMS platforms such as SendGrid, Amazon SES or Telnyx, and deliverability basics (SPF, DKIM, DMARC). Marketing compliance: TCPA consent for texting, CAN-SPAM for email, 10DLC registration. Social media APIs such as LinkedIn, Meta or X. Voice AI or AI calling. Retrieval and embeddings, such as pgvector. Job runners such as Trigger.dev, or queues and event-driven systems. Marketing technology: CRMs, campaign analytics, A/B testing. Product or UX sense for internal tools. How we hire
We aim to go from first conversation to offer very quickly. #20-minute screen with the VP of Engineering. # Interview/assessment on synthetic data. You'll build a small campaign agent that drafts personalized messages, respects an opt-out list and asks for approval before sending, then write an eval that catches its bad outputs. AI tools are allowed; we grade judgment, not typing. # Review conversation walking through your assessment and past projects. # References and background check, then an offer.
~ We hire for potential and drive. If you meet most of the requirements and are hungry to get building and learn the rest, apply immediately.
Our Chief Architect proves what's possible with AI. Your job is to turn those prototypes into reliable products that run every day, measure their results, and stay inside the rules that govern financial communications. This role suits someone early in their career who wants to prove themselves, or an experienced engineer who wants to start fast. You'll get hands-on mentoring, real ownership from week one, and a clear path to grow. Term Detail Type 13-week contract, with the opportunity to extend or convert to full time Rate $30 to $40 an hour, based on experience Hours Full time, 40 hours a week Start As soon as possible Location San Diego, CA, on-site Reports to VP of Engineering and AI Product Development What you'll do Build AI campaign products. Ship AI-written email, SMS and social media campaigns: audience selection, personalized drafts, scheduling, sending and reply handling. Build agents with guardrails. Design LLM agents with tool calling, structured outputs and clear limits on what they can do. Every bulk send, data change or external action goes through a human approval step. Build compliance in, not on. Respect consent, opt-outs, quiet hours and suppression lists for every message. Keep campaign content within the rules for financial communications. Measure quality with evals. Build evaluation sets for each agent, test prompt and model changes against them before release, and track accuracy, tone and policy violations. Connect to our platforms. Integrate with our outbound sending system, our CRM, email and SMS providers, and social media APIs. Track results. Report deliverability, replies, conversions and opt-outs, and use them to improve campaigns. Run AI in production. Log agent runs, monitor cost and latency, handle model failures and fallbacks, and keep secrets and personal data out of prompts and logs. Prototype new ideas. Turn the team's new AI product ideas into working, reviewed prototypes quickly. Review agent-written code. Much of our code is written with AI coding tools. You'll review and test what they produce in your area before it ships. What you need ~1 to 4 years of software engineering experience, including internships, contract work or substantial projects. ~ LLM applications: you've shipped something real on Claude, OpenAI or similar models: prompt design, tool calling, structured outputs and handling model failures. ~ Agentic systems: experience with agent frameworks or patterns such as the Vercel AI SDK, LangGraph, the OpenAI Agents SDK or MCP servers, and an understanding of where agents go wrong. ~ Evals: you know how to build test sets for model output and measure whether a change made it better or worse. ~ Programming: TypeScript (our main language) and some Python; Node.js and React or Next.js. ~ APIs and integrations: REST APIs, webhooks, OAuth and background jobs that retry safely. ~ Data: solid SQL with PostgreSQL; comfortable reading and shaping data for prompts and retrieval. ~ Cloud: working knowledge of AWS or Vercel and serverless deployment. ~ Git and code review: pull requests, CI and tests for the code you write. ~ AI-assisted development: you use tools like Claude Code or Cursor daily, and you catch their mistakes. ~ Judgment: you treat customer data and outbound messages with care, and you know when an agent should stop and ask a person.
Nice to have Fintech: investor communications, payments, KYC, or work under SEC or FINRA rules. Email or SMS platforms such as SendGrid, Amazon SES or Telnyx, and deliverability basics (SPF, DKIM, DMARC). Marketing compliance: TCPA consent for texting, CAN-SPAM for email, 10DLC registration. Social media APIs such as LinkedIn, Meta or X. Voice AI or AI calling. Retrieval and embeddings, such as pgvector. Job runners such as Trigger.dev, or queues and event-driven systems. Marketing technology: CRMs, campaign analytics, A/B testing. Product or UX sense for internal tools. How we hire
We aim to go from first conversation to offer very quickly. #20-minute screen with the VP of Engineering. # Interview/assessment on synthetic data. You'll build a small campaign agent that drafts personalized messages, respects an opt-out list and asks for approval before sending, then write an eval that catches its bad outputs. AI tools are allowed; we grade judgment, not typing. # Review conversation walking through your assessment and past projects. # References and background check, then an offer.
~ We hire for potential and drive. If you meet most of the requirements and are hungry to get building and learn the rest, apply immediately.