Senior AI/ML Engineer - SLM Optimization & Inference

Concentrix; United States; Remote; USD 145,000 - 171,000; Full time

Benefits offered by this job

Medical, dental, and vision insurance

401(k) retirement plan

Paid time off and holidays

Paid learning days

Job summary

Concentrix is seeking a hands-on Senior AI/ML Engineer to fine-tune, optimize, and deploy Small Language Models for production-grade inference, with a focus on Alibaba Qwen. The role emphasizes improving latency, throughput, and GPU efficiency in real-time audio intelligence pipelines.

You will work close to model, infrastructure, and performance benchmarks, building scalable ML systems ready for production and collaborating across engineering and data teams.

Qualifications

• Strong hands-on experience with PyTorch, Hugging Face Transformers, and SLM fine-tuning techniques such as LoRA, QLoRA, and PEFT.
• Practical experience using vLLM or SGLang for high-performance model inference.
• Deep understanding of GPU utilization, CUDA fundamentals, KV caching, and memory optimization.
• Experience evaluating and improving model precision, recall, latency, throughput, and inference performance.
• Strong Python programming and ML engineering skills.
• Proven ability to benchmark, debug, and optimize models for production environments.
• Experience fine-tuning and deploying Alibaba Qwen models is preferred.
• Experience with real-time audio AI, speech recognition, ASR, or audio classification is preferred.
• Experience optimizing AI models for high-throughput, low-latency production environments is preferred.
• Ability to work hands-on across both SLM model training/fine-tuning and GPU inference optimization.
• Strong collaboration skills with the ability to partner effectively across engineering and data teams.

Responsibilities

• Fine-tune and optimize Small Language Models, particularly Alibaba Qwen models, for domain-specific use cases.
• Improve model performance across key metrics, including accuracy, precision, recall, latency, and inference throughput.
• Optimize model serving using vLLM, SGLang, or similar high-performance inference engines.
• Implement advanced inference optimization techniques, including GPU memory optimization, KV-cache management, quantization, and efficient batching.
• Develop, enhance, and optimize real-time audio intelligence pipelines for low-latency processing.
• Benchmark, profile, and tune models for production-scale deployment.
• Collaborate with engineering and data teams to evaluate model behavior, identify bottlenecks, and improve end-to-end ML system performance.
• Build reliable, scalable ML workflows that support high-throughput, low-latency AI applications.
• Apply senior-level technical judgment to balance model quality, infrastructure efficiency, and production performance requirements.

Skills

PyTorch

Hugging Face Transformers

LoRA

QLoRA

PEFT

Python

ML Engineering

GPU Inference

vLLM

SGLang

Tools

vLLM

SGLang

CUDA

KV caching

Memory optimization

Job description

Concentrix is seeking a hands-on Senior AI/ML Engineer to fine-tune, optimize, and deploy Small Language Models for production-grade inference, with a focus on Alibaba Qwen. The role emphasizes improving latency, throughput, and GPU efficiency in real-time audio intelligence pipelines.

You will work close to model, infrastructure, and performance benchmarks, building scalable ML systems ready for production and collaborating across engineering and data teams.