Machine Learning Engineer

About the role

You are your own worst critic and you maintain a high bar for quality, refusing to rest until the job is done right without settling for 90 percent. You ship fast with high agency, actively diving in to fix issues rather than merely voicing them. You own the end to end responsibility for training and deploying new state of the art models that parse and interpret unstructured data within production systems. You experiment with novel techniques to push LLM accuracy forward while building robust data pipelines and rigorously evaluating model performance. You work directly with founders and customers to shape product direction and engineering strategy, translating their needs into technical execution.

Key facts

What you'll do

Train and deploy new state of the art models for parsing and interpreting unstructured data at scale within production environments.

Experiment with novel techniques to improve LLM accuracy, testing hypotheses quickly and iterating based on empirical results.

Build data pipelines that are reliable and efficient, ensuring smooth flow and transformation of document based information.

Evaluate model performance using rigorous quantitative methods, identifying edge cases and driving targeted improvements.

Integrate models into the product seamlessly, ensuring that the user experience benefits directly from advances in ML research.

Work directly with the founders and customers to gather requirements and shape the product direction and engineering strategy.

Break document layouts into subsections and contextually parse each depending on the type of content using a combination of vision models and LLMs.

Leverage a suite of heuristics built in house to handle complex document structures that traditional systems cannot reliably process.

Requirements

You have 2 plus years of experience with training, fine tuning, and evaluating ML models used in production systems, demonstrating a track record of responsible deployment.

You are exceptional at Python or similar languages, writing clean, maintainable code that can scale within production constraints.

You are well versed with both traditional computer vision techniques and modern vision language models, understanding their respective strengths and limitations.

You build your own tools as needed, such as a quick Streamlit app to test hypotheses or create a dataset that accelerates experimentation.

You adopt a quantitative approach to building products, using data to guide decisions and validate hypotheses.

You are comfortable debugging issues at multiple levels, from data quality to model behavior, and iterating rapidly to resolve them.

You get hands on with the full development lifecycle, moving confidently from ideation to implementation and ultimately to shipping users.

You thrive in a fast moving environment where responsibilities shift quickly and ownership is expected across the stack.

Practical notes

This is an in person role at our office in SF.

We're an early stage company which means that the role requires working hard and moving quickly.

Please only apply if that excites you.

Reducto is the agentic document platform for leading AI teams who demand enterprise performance at scale. We provide a comprehensive toolkit for working with documents the way a human would, combining custom in house and leading frontier models to power efficient and accurate document workflows.

We've grown rapidly, increasing revenue 8x year over year and partnering with hundreds of companies, from leading AI teams like Harvey, Vanta, and Scale, to enterprise customers across FAANG and top trading firms.

Reducto has raised over $100M from world class investors including a16z, Benchmark, and First Round Capital.

• Philosophy: You are your own worst critic. You have a high bar for quality and don't rest until the job is done right - no settling for 90%. We want someone who ships fast, with high agency, and who doesn't just voice problems but actively jumps in to fix them.
• Experience: You have 2+ years of experience with training, fine tuning, and evaluating ML models used in production systems
• Language/Skills: You're exceptional at Python or similar, and are well versed with both traditional computer vision and VLMs
• Tools: Build your own tools as needed - like a quick Streamlit app to test hypotheses or create a dataset.
• Approach: A quantitative approach to building products. Ability to debug, experiment, and iterate fast. You should be comfortable getting hands on with the full development lifecycle, from ideation to shipping to users.
• Training and deploying new state of the art models for parsing and interpreting unstructured data
• Experimenting with novel techniques to improve LLM accuracy
• Build data pipelines, evaluate model performance, and integrate models into the product
• Working directly with the founders and customers to shape the product direction and engineering strategy
• Have prior experience founding a company or building products at early stages
• Are ambitious and driven, and care a lot about doing great work with great people
• Keep up with the latest developments in ML/AI

Reducto is an Equal Opportunity Employer committed to diversity and inclusion in the workplace. All qualified applicants will receive consideration for employment without regard to sex, race, color, age, national origin, religion, physical and mental disability, genetic information, marital status, sexual orientation, gender identity or expression, or any other characteristic protected by law.

Nearly 80 percent of enterprise data is in unstructured formats like PDFs. PDFs are the status quo for enterprise knowledge in nearly every industry. Insurance claims, financial statements, invoices, and health records are all stored in a structure that is simply impractical for use in digital workflows. This is not an inconvenience - it is a critical bottleneck that leads to dozens of wasted hours every week https://www.reducto.ai/blog/the-real-cost-of-manual-document-processing.

Traditional approaches fail at reliably extracting information in complex PDFs. OCR and even more sophisticated ML approaches work for simple text documents but are unreliable for anything more complex. Text from different columns are jumbled together, figures are ignored, and tables are a nightmare to get right. Overcoming this usually requires a large engineering effort dedicated to building specialized pipelines for every document type you work with.

Reducto breaks document layouts into subsections and then contextually parses each depending on the type of content. This is made possible by a combination of vision models, LLMs, and a suite of heuristics we built over time. Put simply, we can help you:

• Accurately extract text and tables even with nonstandard layouts
• Automatically convert graphs to tabular data and summarize images in documents
• Extract important fields from complex forms with simple, natural language instructions
• Build powerful retrieval pipelines using Reducto's document metadata
• Intelligently chunk information using the document's layout data

At Reducto, we are invested in the well being and growth of our team. Here is what we currently offer:

• Unlimited PTO: We believe great work requires recharging.
• Lunch: Receive a free lunch to eat with your teammates daily at the office
• Reimbursed Transportation: Provide us with your receipts and we will take care of the costs
• Insurance: Generous health insurance covering medical, dental, and vision.
• Health and Wellness Budget: We provide up to 150 dollars per month reimbursement for health and wellness spending, such as gym memberships, fitness classes, or similar.
• Parental Leave: Work with us to build a leave schedule that works for you and your family.

Reducto is an Equal Opportunity Employer committed to diversity and inclusion in the workplace. All qualified applicants will receive consideration for employment without regard to sex, race, color, age, national origin, religion, physical and mental disability, genetic information, marital status, sexual orientation, gender identity, and expression.