Job Description
Salary: $25 - 30/ hr
\n
About Akoncagua AI
\n
Akoncagua AI builds domain-specialized AI systems for technically demanding industries. Our work includes LLMs, multimodal AI, agentic automation, RAG, evaluation, and production AI infrastructure. We are seeking an AI Automation Engineer Intern to help build secure, reliable AI agents and automated workflows for accounting and enterprise finance.
\n
What you'll work on
\n
- \n
- Build and improve LLM agent workflows for document intake, classification, and transaction coding
- Develop document-intelligence pipelines for invoices, receipts, contracts, and financial statements
- Implement retrieval over accounting policies, vendor records, and historical transactions
- Write evaluation harnesses that measure extraction accuracy, classification quality, and end-to-end workflow outcomes
- Add tracing and observability for prompts, tool calls, failures, latency, and cost
- Integrate agents with accounting platforms, databases, and internal APIs
\n
\n
\n
\n
\n
\n
\n
Required qualifications
\n
- \n
- Currently enrolled in a Master's or PhD program in Computer Science, AI/ML, or a related technical field, or equivalent practical experience
- Strong Python skills, plus solid fundamentals in data structures, debugging, and testing
- Experience building LLM-based applications, whether through coursework, research, personal projects, or internships
- Familiarity with at least one agent framework such as LangGraph, AutoGen/AG2, Semantic Kernel, or the OpenAI Agents SDK
- Working knowledge of RAG and vector databases
- Comfortable with Docker and version control in a collaborative codebase
\n
\n
\n
\n
\n
\n
\n
Preferred qualifications
\n
- \n
- Experience with MCP or similar tool-integration standards
- Experience with n8n, Zapier, Make, or event-driven automation
- Exposure to FastAPI, PostgreSQL, pgvector, or Kubernetes
- Interest in or exposure to accounting concepts such as double-entry bookkeeping, journal entries, reconciliations, or approval controls
- Awareness of prompt-injection risks, least-privilege design, and secrets management
- Open-source contributions or publications at AI conferences
- Prior work with financial platforms, databases, and transactions
\n
\n
\n
\n
\n
\n
\n
\n
How we work
\n
We care about engineers who understand and can defend the design decisions in their code. AI coding tools are welcome here, but you should be able to explain why a system works the way it does and where it breaks.