Staff Engineer, Core ML

W3Global; Austin (TX); On-site; USD 140,000 - 220,000; Full time

Benefits offered by this job

Competitive compensation

Equity

Employer-paid benefits

Flexible PTO

Access to AI tools

Job summary

W3Global is seeking a seasoned ML engineer to ship production-ready models and build robust document-understanding systems for complex technical content including PDFs/CAD. You will advance agent reliability, apply new research to production, and mentor teammates while shaping end-to-end ML pipelines and evaluation strategies.

The role emphasizes hands-on development, data curation, and strong engineering discipline to ensure deterministic, trustworthy AI systems used across diverse document

Qualifications

• Hands-on AI/ML engineering experience with a proven track record of shipping models into production.

Responsibilities

• Develop algorithms that extract meaning from complex technical documents containing text, vector geometry, CAD, and raster content, spanning vision models, geometric algorithms, and language understanding.
• Build document-understanding capabilities into the tools that power long-running agentic systems.
• Improve the reliability, determinism, and effectiveness of agentic systems, including agent planning, tool usage, and failure recovery.
• Build and extend training and evaluation pipelines for core models and agents, ensuring every change is measured against real customer data before deployment.
• Analyze algorithm performance in production, identify failure patterns, and develop targeted improvements.
• Identify gaps and bias in datasets and address them through data curation, labeling, or synthetic data generation.
• Monitor emerging research and techniques and adapt promising approaches for production use.
• Mentor and provide technical guidance to other engineers on the team.

Skills

ML engineering

Production ML

Python

PyTorch

AI Observability

Go

Data engineering

Langfuse

Tools

Langfuse

Opik

vLLM

PyTorch

Job description

Responsibilities

• Develop algorithms that extract meaning from complex technical documents containing text, vector geometry, CAD, and raster content, spanning vision models, geometric algorithms, and language understanding.
• Build document-understanding capabilities into the tools that power long-running agentic systems.
• Improve the reliability, determinism, and effectiveness of agentic systems, including agent planning, tool usage, and failure recovery.
• Build and extend training and evaluation pipelines for core models and agents, ensuring every change is measured against real customer data before deployment.
• Analyze algorithm performance in production, identify failure patterns, and develop targeted improvements.
• Identify gaps and bias in datasets and address them through data curation, labeling, or synthetic data generation.
• Monitor emerging research and techniques and adapt promising approaches for production use.
• Mentor and provide technical guidance to other engineers on the team.

Key Responsibilities

• Develop algorithms that extract meaning from complex technical documents containing text, vector geometry, CAD, and raster content, spanning vision models, geometric algorithms, and language understanding.
• Build document-understanding capabilities into the tools that power long-running agentic systems.
• Improve the reliability, determinism, and effectiveness of agentic systems, including agent planning, tool usage, and failure recovery.
• Build and extend training and evaluation pipelines for core models and agents, ensuring every change is measured against real customer data before deployment.
• Analyze algorithm performance in production, identify failure patterns, and develop targeted improvements.
• Identify gaps and bias in datasets and address them through data curation, labeling, or synthetic data generation.
• Monitor emerging research and techniques and adapt promising approaches for production use.
• Mentor and provide technical guidance to other engineers on the team.

Examples of Work You Could Own

• Fix a Failure Class: Investigate why extraction fails on specific types of technical drawings, such as dense electrical schematics, and improve performance through better data, models, or algorithms.
• Fine-Tune a Specialized Model: Fine-tune a LoRA on an in-house model to extract specialized artifacts from high-resolution blueprint pages.
• Bring a New Algorithm to Production: Evaluate approaches from recent research, such as layout or table-structure models, demonstrate improvements against existing solutions, and make them reliable and efficient enough for production.
• Improve Retrieval: Enhance retrieval quality, relevance, and latency across different document types and customers.
• Make Agents More Deterministic: Investigate why the same task produces inconsistent results across multiple runs and improve the tools or agent loop to deliver consistent outcomes.

Experience

We're looking for someone who has successfully shipped ML-powered software into production, can adapt proven approaches, evaluate alternatives, and measure the impact of improvements on customers.

Ideally, you'll have deep expertise and strong technical opinions in at least one of the following areas:

• Agent harness design and execution.
• Document extraction and structured data understanding.
• High-resolution computer vision.
• Vision-language model (VLM) fine-tuning and inference.
• Retrieval systems, including embeddings and reranking.
• Agent evaluation and optimization.

Required Qualifications

• Hands-on AI/ML engineering experience with a proven track record of shipping models into production.
• Experience training and fine-tuning production models, including building the datasets and pipelines behind them.
• Experience developing and operating ML evaluation pipelines, including datasets, metrics, and regression testing for production models and agents.
• Experience analyzing data distributions, identifying gaps and bias, and curating high-quality training and evaluation datasets.
• Experience with vision and multimodal models, particularly technical documents such as PDFs and CAD files containing text, vector geometry, and raster content.
• Strong knowledge of data structures, algorithms, and software engineering principles.
• Proven expertise in LLMs, deep learning systems, and applied machine learning.
• Experience debugging production ML systems using AI observability tools such as Langfuse, Opik, or similar platforms.
• Proficiency in Python and relevant ML frameworks such as PyTorch and vLLM.
• Daily, fluent use of AI coding tools, including effective prompting, iterative refinement, and the ability to identify incorrect or unreliable AI-generated code.

Preferred Qualifications

• Knowledge of computational geometry or experience with vector drawing formats such as CAD and vector PDFs.
• Experience with Go programming language.
• Familiarity with construction, civil engineering, manufacturing, or related industries where interpreting complex technical drawings and specifications is essential.

Human Skills

• Ownership mindset with a focus on delivering reliable, production-ready systems.
• Strong product judgment and understanding of how system behavior impacts user experience and trust.
• Ability to rigorously review agent instructions, skills, and documentation to ensure clarity, accuracy, and reliability.
• Curiosity about how systems behave in practice and how customers interact with them.
• Bias toward execution and the ability to turn ambiguous problems into working solutions.
• Strong judgment about production readiness and the willingness to challenge decisions that are not grounded in customer outcomes.
• Clear communication skills, with the ability to explain ML tradeoffs to engineers, FDEs, and product stakeholders.
• Critical evaluation of AI-generated output, including identifying hallucinated APIs, incorrect logic, and subtle bugs.
• Systems-thinking mindset, with attention to correctness, determinism, and failure modes in long-running agents.
• Strong measurement discipline and the ability to distinguish genuine improvements from evaluation noise.
• Ability to raise the technical bar through design reviews, code reviews, and mentoring.
• Low-ego approach with a focus on collaboration and overall team success.

What We Offer

• Opportunity to work with a high-caliber team of experienced manufacturing and AI leaders.
• Opportunity to help build a category-defining platform transforming how manufacturers sell, support, and innovate.
• Competitive compensation, meaningful equity, and employer-paid benefits.
• Flexible paid time off.
• Access to the latest AI tools and technologies.
• A values-driven culture that rewards ownership, curiosity, and execution velocity