Staff Engineer - Model Accuracy Development and Test

Role Overview Qualcomm India Private Limited is hiring a Staff Engineer specialising in Inference Accuracy for its Engineering Group. This role is focused on designing, developing, and validating deep learning model accuracy deployed across large-scale data-centre hardware platforms. You will work at the intersection of AI model evaluation, quantization, and inference pipeline engineering to ensure accurate and reliable model outputs at scale. Key Responsibilities - Define and implement accuracy KPIs across multiple precision modes including FP32, FP16, and INT8 - Develop scalable Python-based accuracy evaluation tools and fully automated testing pipelines - Implement accuracy-preserving optimizations across inference frameworks such as TensorRT, ONNX Runtime, AITemplate, and Triton - Build and maintain automated accuracy evaluation pipelines across ONNX, TensorFlow, and PyTorch frameworks - Develop reusable plugins for preprocessing, post-processing, and metric evaluation workflows - Execute comprehensive accuracy tests for large-scale models including LLMs, vision models, and diffusion models - Validate model accuracy under various quantization and mixed-precision settings - Perform deep architectural accuracy analysis covering layers, attention mechanisms, and parameter configurations - Identify and debug issues related to preprocessing drift, tokenization mismatches, operator fallbacks, and quantization effects - Analyze accuracy differences across hardware targets, firmware versions, and runtime backends - Perform slice-based accuracy analysis across batch sizes, concurrency levels, sequence lengths, and domain shifts - Design and run accuracy recovery experiments including fine-tuning, calibration, and hyperparameter adjustments - Document workflows, maintain dashboards, and publish accuracy results for cross-functional stakeholders Required Qualifications - Bachelor's, Master's, or PhD degree in Engineering, Computer Science, Machine Learning, AI, or a related field - 4 to 10 years of software engineering or related work experience depending on degree level - Minimum 2 years of experience in software or system test engineering including test plan development and automation - Strong background in AI and ML model evaluation, accuracy metrics, and statistical analysis - Solid understanding of model architectures including transformers, CNNs, RNNs, and Mixture of Experts - Hands-on experience with inference runtimes such as TensorRT, ONNX Runtime, and Triton - Deep expertise in quantization techniques including INT8, FP8, INT4, calibration, and QAT workflows - Proficiency in Python and familiarity with ML toolkits including TensorFlow and PyTorch - Experience with data-centre accelerators such as NVIDIA A100, H100, B200, AI100 Ultra, Gaudi, and TPU - Familiarity with distributed deployment systems including Kubernetes and cloud inference services - Knowledge of LLM accuracy evaluation tools such as lm-eval and HELM is an added advantage Why Join Us Qualcomm is a global semiconductor and wireless technology leader offering world-class engineering challenges and an inclusive work environment committed to innovation. Joining this team means working on cutting-edge AI inference systems that power next-generation data-centre and AI platforms at a global scale.

- Strong AI and ML model evaluation and accuracy metrics expertise - Hands-on experience with TensorRT, ONNX Runtime, and Triton inference runtimes - Proficiency in Python for automation and pipeline development - Experience with LLMs, generative AI, and large-scale model accuracy validation - Deep knowledge of quantization techniques including INT8, FP8, INT4, QAT, and mixed-precision workflows - Experience with model graph conversion from PyTorch to ONNX to backend engines - Understanding of transformer, CNN, RNN, and MoE model architectures - Experience with data-centre accelerators such as NVIDIA A100, H100, B200, Gaudi, and TPU - Statistical analysis and visualization skills for accuracy debugging - Familiarity with Kubernetes and cloud-based distributed inference systems - 4 to 10 years of software engineering or software test engineering experience

- Opportunity to work on cutting-edge AI and deep learning inference systems - Inclusive and accessible workplace with equal opportunity employment - Global exposure working with world-class data-centre hardware platforms - Collaborative engineering environment within a leading semiconductor company