Machine Learning Engineer

AI/ML Engineer - Soundfr | Estonia | Hybrid We’re hiring an AI/ML Engineer to join the Soundfr team in Estonia. Soundfr helps people who are deaf or hard of hearing stay aware of important sounds at home. Users record sounds such as their doorbell, home appliances or a baby crying to train a personalised sound recognition model. Our connected microphone detects these sounds, and the app alerts the user. We also provide communication tools such as speech-to-text. The role You’ll work directly with our CTO to develop and improve the ML system behind our app and connected microphone. You’ll build on an existing product, with room to influence how we train models, evaluate improvements and deploy them to devices. A central challenge is making personalised sound detection reliable from a small number of user recordings. That means handling background noise, different rooms, microphone placement and recording quality while keeping missed detections and false alerts under control. What you’ll do - Develop and improve pipelines for training personalised models from user recordings. - Improve data preparation, audio preprocessing and training datasets. - Design evaluations that reflect real use and make improvements measurable. - Investigate detection failures and use the findings to improve models and training. - Optimise and deploy models to mobile and edge devices, balancing detection performance, latency and memory use. - Follow changes through deployment and assess how they perform in practice. What we’re looking for - Strong Python skills and practical experience building ML systems. - Experience with a framework such as TensorFlow or PyTorch. - Experience taking models through data preparation, training, evaluation and deployment. - A systematic approach to experiments, debugging and working with imperfect data. - Solid software engineering skills and the ability to turn experiments into maintainable production code. Experience with audio ML, signal processing, learning from small datasets, model quantisation or on-device inference would be a strong advantage. What we offer - Ownership of meaningful parts of the ML system and influence over technical decisions. - Close collaboration with our CTO and a direct connection between your work and user feedback. - Flexible hybrid work in Estonia. - A competitive salary and employee share options. Interested? Apply through LinkedIn or send your CV to [email protected]. We’d also like to hear briefly about an ML project you’ve built or deployed and your contribution to it. #Hiring #MachineLearning #AudioML #Accessibility #Estonia … mais Pesquise vagas mais rápido com Premium Acesse informações sobre empresas, como prioridades estratégicas, tendências de número de funcionários e muito mais Jonatas e outros milhões de usuários têm contas Premium Reative o Premium Aproveite o suporte 24 horas e cancele quando quiser. Sobre a empresa Soundfr 184 seguidores Seguir Tecnologia, Informação e Internet • 2-10 funcionários • 5 no LinkedIn Soundfr is an AI-powered accessibility platform designed for people that are hard of hearing. Our self-trainable sound recognition system allows users to record and teach the AI to detect the sounds that matter most in their daily lives whether it’s a doorbell, baby crying, or a smoke alarm. Once recognized, alerts are instantly delivered to smartphones and smartwatches, helping users stay safe, independent, and aware. Unlike generic solutions, Soundfr is specially built for the hard of hearing community, ensuring trust, accuracy, and relevance. Backed by over two years of co-creation with users, we combine cutting-edge AI with a subscription-based model that removes cost barriers, offering continuous updates and scalability across markets.