Your Coupa resume, matched to the job ad
Drop in your resume and it opens next to this job, ready to tailor in one click.
4.8 from 167 in-app reviews
Principal Engineer, AI/ML Architecture - 11310
Coupa · Foster City, CA, United States
Tailor your resume to this job
No account needed to start.
What Coupa asks for
From Coupa's Principal Engineer, AI/ML Architecture - 11310 ad, checked 3 Nov 2026.
Model training and deployment
At least 5 years of software engineering focused on ML/AI systems. You must have experience training or fine-tuning large language models and have shipped a domain-adapted or fine-tuned model to production.
ML architecture and infrastructure
Working knowledge of transformer architectures and training optimization techniques like LoRA, QLoRA, and RLHF. Experience with distributed training on GPU clusters and proficiency in Python and PyTorch.
RAG and knowledge retrieval
Designing retrieval-augmented generation systems that integrate vector search, embedding models, and structured knowledge or knowledge graphs.
Technical leadership and mentoring
Directing the technical path for AI capabilities, prototyping new approaches, and guiding a global team of machine learning and data engineers.
The ad asks for 15 or more years of software engineering experience, including at least 5 years focused on ML/AI systems. Based in Foster City, CA.
How the example answers the ad
Model training and deployment
“Trained and fine-tuned Llama-3 models on private financial datasets to automate invoice categorization with 94 percent accuracy.”
Fine-tuning Llama-3 models for invoice categorization shows experience with model training and deployment of a domain-adapted system.
ML architecture and infrastructure
“Designed and deployed distributed training systems for custom NLP models across 500+ AWS GPU instances to process massive financial logs.”
Managing 500+ AWS GPU instances for distributed training of NLP models shows the scale required for ML architecture and infrastructure.
RAG and knowledge retrieval
“Built a RAG pipeline utilizing Pinecone and custom knowledge graphs to enhance spend-management insights for enterprise customers.”
The use of Pinecone and custom knowledge graphs to build a RAG pipeline shows experience with RAG and knowledge retrieval.
Technical leadership and mentoring
“Directed a cross-functional team of 12 engineers, increasing feature delivery velocity by 35 percent through agile MLOps workflows.”
Directing a 12-person cross-functional team to improve delivery speed shows the team guidance part of technical leadership and mentoring.
The whole example as text
Victor Nguyen · Principal AI/ML Architect
Summary
Principal AI/ML Architect with 16 years of experience in high-scale distributed systems and deep learning infrastructure. Architected and deployed domain-adapted Llama models for financial spend management using PyTorch and AWS GPU clusters. Led distributed teams of 12 engineers across North America and Asia to ship production-grade retrieval-augmented generation pipelines. Optimized inference latency by 40 percent while reducing total compute costs by 25 percent through custom quantization and speculative decoding.
Experience
Principal AI/ML Architect · Stripe
San Francisco, CA · 2020 - Present
- Trained and fine-tuned Llama-3 models on private financial datasets to automate invoice categorization with 94 percent accuracy.
- Built a RAG pipeline utilizing Pinecone and custom knowledge graphs to enhance spend-management insights for enterprise customers.
- Directed a cross-functional team of 12 engineers, increasing feature delivery velocity by 35 percent through agile MLOps workflows.
- Reduced LLM inference costs by 28 percent through implementation of vLLM serving and speculative decoding on A100 GPU clusters.
Senior AI Lead · Palantir Technologies
Palo Alto, CA · 2015 - 2020
- Designed and deployed distributed training systems for custom NLP models across 500+ AWS GPU instances to process massive financial logs.
- Developed a proprietary graph-based embedding model that reduced data retrieval latency from 2 seconds to 150 milliseconds.
- Mentored 8 junior data scientists, resulting in 3 internal promotions to senior engineering roles over a 3-year period.
- Integrated PyTorch Lightning pipelines into production CI/CD environments to shorten model deployment cycles from weeks to days.
Machine Learning Engineer · Amazon
Seattle, WA · 2012 - 2015
- Optimized ranking algorithms for recommendation engines by implementing distributed reinforcement learning on internal cloud clusters.
- Led the migration of legacy recommendation services to a microservices-based architecture, maintaining 99.99 percent uptime during transition.
- Developed feature extraction pipelines in Python that processed 10 terabytes of daily transaction data with 100 millisecond latency.
Software Engineer · Cisco
San Jose, CA · 2010 - 2012
- Developed core software components for network management systems using Python and C++, handling 5 million concurrent connections.
- Engineered automated testing suites that identified and resolved 200+ production bugs, improving stability by 40 percent.
- Contributed to the design of scalable telemetry data processing systems that supported real-time performance monitoring.
Education
Master of Science in Computer Science, Specialization in AI · Stanford University
Stanford, CA · 2008 - 2010
- Completed thesis on distributed neural network training optimization techniques.
- Graduated with 3.9 GPA and university honors for outstanding research contribution.
- Served as a teaching assistant for graduate-level machine learning and algorithmic complexity courses.
Bachelor of Science in Electrical Engineering · University of California, Berkeley
Berkeley, CA · 2004 - 2008
- Awarded Dean's List honors for four consecutive semesters.
- Participated in senior capstone project focused on real-time signal processing hardware.
- Contributed to research projects regarding high-performance computing architectures.
Skills
Frequently asked questions
Is the example from an employee?
No. It is an AI-written example for this ad, made to show how a resume can answer it.
Is enterprise software experience required?
It is preferred but not strictly required. The ad states that experience with enterprise B2B SaaS platforms is a preferred qualification.
Where do I apply?
On the official Coupa posting, through the job link on this page.
What happens when I tailor my resume here?
Your resume opens in the editor next to this job. Tailor it in one click, then keep or discard the edits. You only need an account to download.