Senior Generative AI Engineer, LLM Optimization

EPAM Systems; United States; Remote; USD 150.000 - 210.000; Vollzeit

Zusammenfassung

EPAM Systems seeks a Generative AI Engineer to own improving the platform's language models—from fine-tuning to deployment. You will partner with Backend teams handling serving infra and collaborate with Data Analysts to define what a good recommendation looks like.

Responsibilities include data curation, synthetic data generation, and distillation for efficiency, with exposure to Vertex AI and GCS. Strong Python and English skills required.

Qualifikationen

• 5+ years of hands-on experience with Large Language Models (LLM) fine-tuning (SFT/PEFT/LoRA).
• Minimum 1 year of leadership experience.
• Proficiency with Transformers/PEFT and related tooling; experience with Vertex AI or equivalent.
• Solid understanding of training mechanics (tokenisation, packing, learning rate).
• Knowledge of catastrophic forgetting and GPU/TPU memory constraints.
• Expertise in LLM evaluation design and metrics.
• Data curation skills with high-quality dataset practices.
• Sustained LLMOps experience including versioning and drift monitoring.
• Strong Python skills for notebooks, training scripts, and eval pipelines.
• Excellent English communication; RLHF and distillation/quantisation knowledge are a plus.

Aufgaben

• Own model fine-tuning, evaluation, and deployment for retail next-best-action tasks.
• Manage dataset lifecycle, annotation workflows, deduplication, and eval set decontamination.
• Generate and curate synthetic data; version datasets used in training.
• Optimize cost/latency via distillation, quantisation, batching, and caching.
• Package and deploy tuned models on Vertex AI or equivalent; manage versions.
• Monitor drift and quality decay; define retraining triggers and rollback paths.
• Implement safety, PII controls, and output validation for enterprise data.
• Collaborate with commercial teams to translate expectations into training signals.

Kenntnisse

LLM fine-tuning

Leadership

Transformers

PEFT

Vertex AI

GCS

Python

English (written/spoken)

RLHF

Distillation

Quantisation

LLMOps

Model versioning

Drift monitoring

GPU memory optimization

Inference stacks

Security & PII controls

Tools

Hugging Face Transformers

Axolotl

Unsloth

Vertex AI endpoints

Jobbeschreibung

EPAM Systems seeks a Generative AI Engineer to own improving the platform's language models—from fine-tuning to deployment. You will partner with Backend teams handling serving infra and collaborate with Data Analysts to define what a good recommendation looks like.

Responsibilities include data curation, synthetic data generation, and distillation for efficiency, with exposure to Vertex AI and GCS. Strong Python and English skills required.