Senior AI/ML Engineer - SLM Optimization & Inference

Concentrix Experts Colombia S.A.S.; United States; On-site; USD 145,000 - 171,000; Full time

Benefits offered by this job

Medical insurance

Dental insurance

Vision insurance

Employee assistance program

Job summary

Concentrix Experts Colombia S.A.S. seeks a Senior AI/ML Engineer to fine-tune Alibaba Qwen models and deploy production-grade inference with optimized GPU usage. You’ll improve latency, throughput, and model quality while contributing to real-time audio intelligence pipelines.

The role involves hands-on model training, deployment, and collaboration with cross-functional teams to balance performance with infrastructure efficiency. Strong PyTorch/Transformers and PEFT expertise are essential.

Qualifications

• Proficient in PyTorch, Transformers, and PEFT techniques (LoRA/QLoRA).
• Hands-on experience with high-performance model inference frameworks like vLLM or SGLang.

Responsibilities

• Fine-tune and optimize Small Language Models, especially Alibaba Qwen variants.
• Improve metrics: accuracy, precision, recall, latency, throughput.
• Optimize model serving with vLLM/SGLang; manage GPU memory and KV caches.
• Implement advanced inference optimization: quantization, batching, memory usage.
• Develop real-time audio intelligence pipelines for low-latency processing.
• Benchmark, profile, and tune for production-scale deployment.
• Collaborate with engineering/data teams to fix bottlenecks.
• Build scalable ML workflows for high-throughput AI apps.
• Exercise senior-level judgment balancing quality and production needs.

Skills

PyTorch

Hugging Face Transformers

LoRA

QLoRA

PEFT

vLLM

SGLang

CUDA

KV caching

GPU memory optimization

Python

Tools

PyTorch

Hugging Face Transformers

vLLM

SGLang

LoRA

QLoRA

PEFT

CUDA

KV caching

Job description

Concentrix Experts Colombia S.A.S. seeks a Senior AI/ML Engineer to fine-tune Alibaba Qwen models and deploy production-grade inference with optimized GPU usage. You’ll improve latency, throughput, and model quality while contributing to real-time audio intelligence pipelines.

The role involves hands-on model training, deployment, and collaboration with cross-functional teams to balance performance with infrastructure efficiency. Strong PyTorch/Transformers and PEFT expertise are essential.