Note: The job is a remote job and is open to candidates in USA. OLarry is a company focused on transforming the tax landscape with technology. As an Applied AI Engineer, you will play a crucial role in developing and implementing innovative tax preparation and planning solutions.
Responsibilities
- Ensure the timely and high-quality delivery of engineering projects by setting clear priorities and deadlines
- Build a pipeline that takes messy PDFs or scanned tax docs and extracts structured, searchable data — even when the AI gets it half-right
- Fine-tune how we use AI and vector search to automatically classify uploaded documents into categories like “Filed Return” or “Supporting Doc”
- Create a confidence scoring system for LLM outputs so downstream services know when to trust the data, and when to loop in a human
- Prototype an AI-driven “Tax Scenario Planner” where users can explore what happens to their taxes if they sell stock, start a side business, or move states
- Build a module that looks at a customer’s past return data and suggests tax-saving opportunities or gaps, using both rules and AI inference
- Implement a forecasting tool that simulates future tax outcomes based on different income streams and entity structures
- Help design a system that asks smart follow-up questions to refine a scenario, like an AI-powered tax assistant
- Build LLM-powered summarization tools that explain complex return data to non-experts in plain language
- Create internal tools for financial advisors or CPAs to visualize multiple return scenarios side-by-side for their clients
- Explore collaborative filtering approaches to recommend planning strategies based on what similar users have done, without breaching data privacy
Skills
- You've built and shipped real-world AI features — especially ones involving documents, workflows, or user-facing automation
- You're comfortable working with large language models (LLMs) and know how to make them useful, whether that's through prompt design, retrieval-augmented generation (RAG), or smart fallback logic when things go sideways, multi-agent pipelines
- You've worked with vector databases, semantic search, or memory-like systems, and have opinions about how to make them fast and relevant
- You think in systems — how to break down complex, fuzzy tasks (like reading a tax form) into repeatable, reliable steps
- You're hands-on with Python and comfortable owning ML pipelines end-to-end, from prototype to production
- You've worked with both structured (relational tables) and unstructured data (PDFs, forms, notes), and know how to stitch them together in a useful way
- You care about the user experience, not just the model output, and like working closely with product and design
- Previous experience with fintech or similarly regulated space
- A high level of EQ, personal drive, and professionalism that adheres to the OLarry Core Pillars
- Bonus if you've built tools or workflows involving scenario modeling or forecasting (e.g., helping users explore “what-if” outcomes)
- Bonus if you've built tools or workflows involving recommendation systems (especially collaborative filtering or ranking systems)
- Bonus if you've built tools or workflows involving multi-agent coordination or building agents that can reason through complex tasks
Benefits
- Unlimited PTO.
- Stock Options.
- Healthcare Insurance - 100% coverage for the employee, 50% coverage for dependents.
- We are committed to investing in your career by offering continuous training and support, ensuring you acquire new skills and opportunities.
Company Overview