Computer Vision Software Engineer

Kulicke & Soffa; Fort Washington (Montgomery County); On-site; USD 150,000 - 210,000; Full time

Job summary

Kulicke and Soffa in the United States is seeking a senior computer vision and ML software engineer to design, implement, and optimize real-time image-based solutions for semiconductor packaging equipment.

You will work with multi-discipline teams to analyze requirements, develop CNN-based models, and deploy them on resource-constrained edge devices while ensuring low latency and high accuracy.

Qualifications

• Proficient Object-Oriented Analysis & Design Methodologies in C++.
• Deep theoretical & practical background in Applied Mathematics, Optimization, Signal Processing, Object/Pattern Recognition.
• Experience in debug and bring-up of prototype HW/SW systems including optics, illumination, imaging sensors, and their interactions with motion systems.
• Experience with developing real-time C++ image processing algorithms under Linux, TCP/IP Networking, multithreading, and inter-process communication.
• Experience with 3rd party imaging tools/APIs/SDKs such as MIL (Matrox Imaging Library), Matlab, OpenCV or other open source libraries is a plus.
• 5+ years of CNN-based computer vision algorithm and product development experience.
• Experience implementing, training, and evaluating image-based classification, object detection, segmentation, and anomaly detection models.
• Experience deploying machine-learning models at scale in resource-constrained edge environments with ONNX, OpenVINO, and/or TensorRT.
• Deep understanding of convolutional neural networks, including receptive fields, feature hierarchies, residual networks, bottleneck layers, attention mechanisms, optimization behavior, augmentation effects, calibration techniques, and common failure modes.
• Strong background in scientific experimentation, including ablation studies, error analysis, robustness evaluation, statistical validation, and root-cause analysis of model behavior.
• Excellent Python programming skills using PyTorch, pandas, numpy, matplotlib, MLflow, and Optuna.
• Master’s or PhD in Computer Science, Computer Engineering, Software Engineering or related fields.

Responsibilities

• Work closely with multi-site software, systems, and application teams to develop and deliver computer vision and machine-learning solutions for advanced semiconductor packaging equipment.
• Review, analyze, and specify product, system, and customer requirements for vision and ML-based features.
• Design, implement, test, and maintain production computer vision software and algorithms using C++ and Python.
• Develop, train, evaluate, interpret, optimize, and deploy image-based machine-learning models.
• Work with system engineers and domain experts to identify high-value inspection, process optimization, root cause analysis, and equipment automation opportunities that improve machine UPH, accuracy, yield, and MTBA.
• Collaborate with mechanical, electrical, motion, optics, process, and application engineers to assess system design alternatives and understand interactions among imaging, illumination, motion, hardware, and algorithms to resolve customer issues.

Skills

C++

Python

CNN-based CV

OOAD

Multithreading

Linux

ONNX

OpenVINO

TensorRT

OpenCV

Mathematics/Optimization

PyTorch

Education

Master’s or PhD in Computer Science/Engineering

Tools

MIL

Matrox MIL

OpenCV

Job description

At Kulicke and Soffa, (NASDAQ:KLIC) our engineers make that decision daily. We are an international leading provider of robotic semiconductor packaging and electronic assembly solutions supporting the global automotive, consumer, communications, computing and industrial segments. Meaning…medical equipment, cell phones, autonomous automobiles, computers, space operations …all of these use products that our equipment helps to make.

Duties to include:

• Work closely with multi-site software, systems, and application teams to develop and deliver computer vision and machine-learning solutions for advanced semiconductor packaging equipment.
• Review, analyze, and specify product, system, and customer requirements for vision and ML-based features.
• Design, implement, test, and maintain production computer vision software and algorithms using C++ and Python.
• Develop, train, evaluate, interpret, optimize, and deploy image-based machine-learning models.
• Work with system engineers and domain experts to identify high-value inspection, process optimization, root cause analysis, and equipment automation opportunities that improve machine UPH, accuracy, yield, and MTBA.
• Collaborate with mechanical, electrical, motion, optics, process, and application engineers to assess system design alternatives and understand interactions among imaging, illumination, motion, hardware, and algorithms to resolve customer issues.

Experience required:

• Proficient Object-Oriented Analysis & Design Methodologies in C++
• Deep theoretical & practical background in Applied Mathematics, Optimization, Signal Processing, Object/Pattern Recognition.
• Experience in debug and bring-up of prototype HW/SW systems including optics, illumination, imaging sensors, and their interactions with motion systems
• Experience with developing real-time C++ image processing algorithms under Linux, TCP/IP Networking, multithreading, and inter-process communication.
• Experience with 3rd party imaging tools/APIs/SDKs such as MIL (Matrox Imaging Library), Matlab, OpenCV or other open source libraries is a plus.
• 5+ years of CNN-based computer vision algorithm and product development experience.
• Experience implementing, training, and evaluating image-based classification, object detection, segmentation, and anomaly detection models.
• Experience deploying machine-learning models at scale in resource-constrained edge environments with strict latency, memory, and compute constraints, including with the use of ONNX, OpenVINO, and/or TensorRT.
• Deep understanding of convolutional neural networks, including receptive fields, feature hierarchies, residual networks, bottleneck layers, attention mechanisms, optimization behavior, augmentation effects, calibration techniques, and common failure modes.
• Strong background in scientific experimentation, including ablation studies, error analysis, robustness evaluation, statistical validation, and root-cause analysis of model behavior.
• Excellent Python programming skills using PyTorch, pandas, numpy, matplotlib, MLflow, and Optuna.
• Master’s or PhD in Computer Science, Computer Engineering, Software Engineering or related fields.

Preferred skills:

• Experience in the architecture, design, and development of real-time vision and control software for semiconductor inspection, advanced packaging, wire bonding, die bonding, or related manufacturing applications.
• Experience with 3D inspection algorithms, depth reconstruction, photometric stereo, depth-from-focus, structured light, or related 3D vision techniques.
• Experience with mechanistic interpretability, feature visualization, attribution methods, representation analysis, explainable AI, and investigation of learned model behavior.
• Experience with image generation using diffusion methods and 3D modeling using Blender
• Experience with Vision Transformers (ViT), multimodal vision-language models, including fine-tuning with LoRA and prompt engineering.
• Experience with representation learning, embedding analysis, contrastive learning, self-supervised learning, reinforcement learning, reward design, policy optimization, offline reinforcement learning, or agentic AI systems

Interested?

• We began as a start-up and now 70 years later as a successful public corporation, still retain the spirit of offering our core values to employees and customers alike. With R&D centers of excellence in the USA, Europe and Asia, we bring together today’s generation of talent, collaborating on tomorrow’s technologies. For more information,