Looking for a job

Ruslan User

@ruslanuser7

Message

Senior AI Backend Engineer at Hugging Face who cut model storage by 70% and GPU costs by 50%.

United States

Message

At Hugging Face, I architect production LLM inference infrastructure and build model-export and compilation pipelines for enterprise AI workloads. I compressed transformer models to INT8, reducing storage requirements by approximately 70% while maintaining production model quality.

I’ve optimized inference across CPU, GPU, and accelerator environments using ONNX Runtime, Optimum, OpenVINO, and AWS Inferentia. These improvements increased inference throughput by approximately 30% and reduced GPU infrastructure costs by approximately 50%.

Earlier, I co-developed secure enterprise search infrastructure at Glean and a real-time payment settlement platform at Sphere. At Accenture, I mentored 16 developers while modernizing banking systems and guiding backend architecture decisions.

Experience

Work history, roles, and key accomplishments

Current

Senior AI Backend Engineer

Current

Hugging Face

Apr 2022 - Present (4 years 7 months)

Architected production LLM inference infrastructure and optimized model serving with quantization and hardware-specific execution. Reduced model storage by 70% and GPU costs by 50% while improving throughput and latency.

Python PyTorch fastAPI gRPC Hugging Face Transformers ONNX Runtime Optimum INT8 IO Binding OpenVINO AWS Inferentia Kubernetes Docker Prometheus OpenTelemetry

AI Backend Engineer

Labelbox

Jun 2021 - Mar 2022 (9 months)

Built model-assisted labeling infrastructure and high-throughput ETL pipelines. Reduced labeling time by 70% and annotation costs by over 40%.

Python fastAPI Celery ETL Webhooks Asynchronous Processing NDJSON

Senior Backend Engineer, AI Applications

Glean

Aug 2019 - Jun 2021 (1 year 10 months)

Co-developed secure multi-tenant enterprise search infrastructure with hybrid retrieval and permission-aware access. Scaled platform to 40+ enterprise pilots with sub-200 ms search performance.

Go Java Kafka Redis Elasticsearch Vector Search Semantic Search ACLs Multi tenant Isolation Kubernetes

Senior Backend Engineer

Sphere

Mar 2018 - Jul 2019 (1 year 4 months)

Co-developed a real-time payments settlement platform, replacing nightly batch processing. Achieved 95% settlement accuracy and reduced latency from 24 hours to under 200 ms.

Go Java Kafka Redis PostgreSQL Distributed Systems Event Driven Architecture Idempotency Rate limiting

Technical Lead, Backend Systems

Accenture

Sep 2017 - Mar 2018 (6 months)

Modernized a high-volume banking platform by decomposing monoliths into microservices and leading migration to AWS and Kubernetes. Mentored 16 developers and reduced API latency by 40%.

Java Spring Boot Microservices AWS Kubernetes OAuth OpenID Connect CI CD Mentorship

Senior Backend Developer

Accenture

May 2017 - Sep 2017 (4 months)

Built Java/Spring Boot microservices and REST APIs for banking workflows, introducing Kafka-based async processing and Redis caching. Reduced database load by 55%.

Java Spring Boot Kafka Redis REST APIs Microservices Idempotency Distributed Locks

Backend Developer

Accenture

Jan 2016 - May 2017 (1 year 4 months)

Developed Java/Spring Boot microservices for account management and transaction history, with Hibernate/JPA and relational databases. Achieved over 90% test coverage.

Java Spring Boot Hibernate JPA REST APIs Junit Mockito CI CD