AI软件工程师
AI Software Engineer
Description
关于职位：
我们的研发团队为全球零售行业构建前沿的视觉智能，分析全球主要市场中的复杂店内场景和密集货架环境。我们结合深度学习、现代视觉-语言模型（VLM）和生成式AI，从大规模零售图像中提取实时理解。
最近，我们扩展到增强现实（AR）领域——将机器学习、2D/3D几何学和计算机图形学相结合，直接在边缘设备上提供交互式、空间化的体验。作为该团队的AI软件工程师，你将专注于解决非平凡的算法挑战，开发视觉和多模态架构，并部署生产级解决方案，覆盖全球数千家门店。至关重要的是，这是一个以工程为主的职位：大约80%的时间将用于编写高质量、完全测试的生产代码，而不是进行纯粹的学术研究。
关于FORM
FORM为全球二十亿移动工作者提供服务，帮助他们彻底改变公司和行业，通过提升一线执行效率的移动技术。FORM的解决方案包括AI驱动的任务管理平台GoSpotCheck、Trax的图像识别技术以及FORM OpX，所有这些都能激活并连接现场团队——与领导者、任务和彼此——使他们能够在企业中取得成功。拥有超过25年的经验，FORM支持来自45个国家一些世界上最知名品牌的10万+全球用户。
要求
关键职责：
- 软件工程与代码质量：编写干净、可维护且完全测试的代码。积极参与代码审查，并使用敏捷方法（Jira）和版本控制（Git）与团队紧密协作。
- 模型开发与优化：训练和评估针对细粒度零售场景理解和产品识别的深度学习模型和视觉-语言模型（VLM）。
- AR与空间解决方案：设计和实现连接计算机视觉和计算机图形学的功能，应用2D/3D几何原理来推动交互式增强现实工作流程。
- 边缘与设备端部署：对非平凡算法进行性能分析、优化和量化，以适应延迟和内存受限的边缘硬件（例如移动设备和嵌入式系统）。
- VLM与生成式AI创新：实际操作多模态大语言模型（LLMs）、视觉-语言模型（VLM）和高级提示技术
查看英文原文
Description
About the Role:
Our R&D team builds cutting-edge visual intelligence for the global retail industry, analyzing complex in-store scenes and dense shelf environments across major markets worldwide. We combine deep learning, modern Vision-Language Models (VLMs), and generative AI to extract real-time understanding from retail imagery at massive scale.
Recently, we expanded into Augmented Reality (AR) - merging machine learning, 2D/3D geometry, and computer graphics to deliver interactive, spatial experiences directly onto edge devices. As an AI Software Engineer on this team, you will focus on solving non-trivial algorithmic challenges, developing visual and multimodal architectures, and shipping production-grade solutions deployed across thousands of stores globally. Crucially, this is an engineering-first role: roughly 80% of your time will be spent writing high-quality, fully tested production code rather than conducting pure academic research.
About FORM
FORM powers the world’s two billion mobile workers as they change companies and industries for good, with mobile technology that improves execution from the frontline. FORM solutions include the AI-enabled task management platform GoSpotCheck, Trax’s image recognition technology, and FORM OpX, all of which activate and connect teams in the field – with leaders, missions, and each other – so they can deliver success in the enterprise. With more than 25 years of experience, FORM supports 100,000+ global users from some of the world’s more recognizable global brands in 45 countries.
Requirements
Key Responsibilities:
- Software Engineering & Code Quality: Write clean, maintainable, and fully-tested code. Actively participate in code reviews and collaborate closely with the team using Agile methodologies (Jira) and version control (Git).
- Model Development & Optimization: Train and evaluate deep learning models and Vision-Language Models (VLMs) tailored for fine-grained retail scene understanding and product recognition.
- AR & Spatial Solutions: Design and implement features bridging computer vision and computer graphics, applying 2D/3D geometric principles to power interactive Augmented Reality workflows.
- Edge & On-Device Deployment: Profile, optimize, and quantize non-trivial algorithms for latency- and memory-constrained edge hardware (e.g., mobile devices and embedded systems).
- VLM & Generative AI Innovation: Work hands-on with multimodal LLMs, VLMs, and advanced prompting techniques to create next-generation automated recognition tools.
- Algorithm Refinement & Scaling: Enhance and optimize existing algorithms to ensure high efficiency, robustness against varying store conditions, and global scalability through rigorous software architecture.
Requirements & Qualifications:
- Experience: 5+ years of hands-on experience developing and deploying Machine Learning and Deep Learning models in production or applied R&D environments.
- Computer Vision & Math Foundations: Strong understanding of classical and modern computer vision, along with the linear algebra, calculus, and mathematical foundations required for 2D/3D geometry and spatial graphics.
- Multimodal & Generative AI: Practical experience working with modern LLMs/VLMs and prompt engineering techniques.
- Core Tech Stack: High proficiency in Python, deep learning frameworks (Keras, Tensorflow, Jax preferred), and standard software development tools.
- Research-to-Product Mindset: Strong problem-solving ability to take complex academic/R&D concepts and adapt them into robust, non-trivial, and highly maintainable production features.
- Communication: Comfortable collaborating daily in English, both written and verbal, in a distributed international team.
- Software Engineering Best Practices: Strong commitment to writing clean code, unit/integration tests, and using Git. Experience working within a structured team environment (code reviews, Jira, CI/CD pipelines).
Nice to Have:
- Experience with Augmented Reality frameworks (e.g., ARKit, ARCore) or computer graphics pipelines (OpenGL, WebGL, or game engines).
- Edge Computing: Familiarity with optimizing models for edge devices and inference runtimes (e.g., CoreML, or TFLite).
- Prior domain experience in retail analytics, planogram compliance, or fine-grained object recognition.
What We Offer:
- Full remote or hybrid work from our cozy office in Krakow.
- The option to enroll in a medical package on preferential terms.
- Subsidized English language courses.
- Company-provided work equipment.
- Opportunities for internal growth and career development.
- Access to internal learning resources.
Compensation: 28,000 - 35,000 PLN per month. Exact compensation may vary depending on skills, experience, and location within Poland.
Originally posted on Himalayas