About the Role:

You’ll design AI systems that sit inside client workflows—turning accounting data and raw reports into clear, human-like messages and summaries using large language models (LLMs). This includes prompt engineering, AI-driven parsing, and embedding simple AI decision-making into our platform.

Responsibilities:

• Build prompts that turn structured/unstructured accounting data into readable summaries
• Fine-tune or instruct LLMs to generate client-friendly messages (e.g., invoice reminders)
• Apply retrieval-augmented generation (RAG) for document-based workflows
• Work with embeddings and document parsing (e.g., PDF annual reports)
• Collaborate with backend team to embed AI in automation flows

Requirements:

• 3+ years in applied AI/NLP or LLM development
• Strong experience with OpenAI APIs, LangChain, or LlamaIndex
• Good knowledge of Python and hands-on experience with core ML frameworks and libraries (e.g., TensorFlow, PyTorch, Scikit-learn, Hugging Face).
• Familiarity with RAG pipelines, vector stores, and data parsing (PDF, HTML)
• Experience in building AI-driven decision-making or summarization flows
• Solid Fundamentals: A strong understanding of machine learning algorithms, data structures, and software engineering best practices.
• Problem-Solving Mindset: A proactive and creative approach to tackling complex challenges.
• Bonus Points: Experience building agentic AI products (AI agents that take initiative, reason, and act) and owning products end-to-end from architecture and prototyping to deployment and optimization.