# Careers Source: https://www.speechmatics.com/company/careers Learn more about working at Speechmatics. ## Join the team on a mission to understand every voice Resources - Careers Orange Push the boundaries of Voice and Conversational AI at Speechmatics and shape the future of communication. - [Open positions](https://www.speechmatics.com/company/careers/roles) ## Life at Speechmatics Jade ### undefined - Column 2 Jade ### **Who we are** At Speechmatics Orange Paving the way with Voice AI We are an inclusive and dynamic team where **innovation thrives**, solving real-world problems with cutting edge speech technology. Encouraging growth, creativity, and collaboration, **together we promote a people-first approach**, supporting personal and professional development. ## Working with us ### **Hybrid work, connected team ** Flexible working Amber Balancing flexibility with collaboration A hybrid approach with 2-3 designated office days each week at our [Cambridge, London, and New York locations](https://www.speechmatics.com/company/contact). - [See our locations](https://www.speechmatics.com/company/contact) ## **Our values** Discover a place where your contributions matter, where innovation is the norm, and where you’re part of something greater. ### People First Jade We believe in putting people first. We don’t just want the best teams; we want the best values fit. When you’re **part of our team**, we’ll do all we can to help you develop your skills. ### Care Deeply Amber Our aim is to ‘Understand Every Voice’. This thinking doesn’t just extend to our technology. We **care deeply about our customers**, especially when it comes to the impact our actions have on the world. ### Be Ambitious Orange Making positive change within our industry takes **bold and forward-thinking action**. Which is why we believe in big goals. Our results show that while change isn't easy, it’s always worthwhile. ### Move Fast Jade We believe in finding the perfect balance between the complex and the simple. When you become part of a collaborative team like ours, you have the **freedom to fail fast and move on**. ## **Perks of working with us** Jade We’re committed to supporting our team with a range of benefits to enhance your well-being, professional growth, and work-life balance. _Please note, some benefits vary by location._ ### Financial Jade Competitive Salary Income Protection Independent Financial Advice Hybrid Working Allowance Generous Pension Scheme Referral Bonus Holiday buy and sell Patent Bonus ### Health Jade Private Medical Insurance Private Dental Care Life Insurance Massages Cycle to work scheme Yoga [Hertility Health](https://hertilityhealth.com/) Bike doctor ### Wellbeing Jade Flexible working Enhanced parental leave Learning & Development Budget Work from abroad (four weeks per year) EV scheme Choice of home office equipment Free lunches Give as you earn & charity day ## You belong at Speechmatics ### **Your voice matters** Inclusivity Amber Inclusivity drives our innovation At Speechmatics, our mission is simple: **Understand Every Voice**. And that goes beyond our technology - it’s the core of who we are. We embrace and celebrate diverse experiences, perspectives, and identities. Our team thrives in an inclusive environment where everyone is supported and valued, regardless of identity, race, disability, age, sexual orientation or religion. - [Open positions](https://www.speechmatics.com/company/careers/roles) ## **Our Peoples' Journey’s** ### Eloy's story Machine Learning Internship - [From intern to innovator](https://www.speechmatics.com/careers/eloy-de-jong-story) #### **From Intern to Innovator: A Journey at Speechmatics** Eloy shares his experience transitioning from academia to industry, working on impactful machine learning projects, and finding the perfect balance of support and independence. Learn how mentorship, collaboration, and hands-on work at Speechmatics have prepared him for a successful career in AI. Amber #### Eloy de Jong Machine Learning Engineer https://www.linkedin.com/in/eloy-de-jong-89b544109/ ### Farah's story Data Engineer - [Farah's story](https://www.speechmatics.com/careers/finding-my-voice) #### **Discover how Speechmatics fosters inclusivity and growth** Farah reflects on her journey, from being drawn to the company's mission of "understanding every voice" to the support she’s received from leadership and peers. Learn how collaboration, flexibility, and a positive environment for women in tech have shaped her experience and accelerated her career development. Orange #### Farah Gouda Data Engineer https://www.linkedin.com/in/farah-gouda/ ### Teri's story Director of Engineering - [Teri's story](https://www.speechmatics.com/careers/teri-drummond-on-inclusive-tech-ai-leadership-and-the-power-of-flow) #### **Building inclusive AI and leading with impact** We sat down with Teri Drummond, our Director of Engineering, to hear more about her career journey to date, what excites her most about working at Speechmatics, and how she’s helping shape the future of responsible, inclusive AI. Read on to discover what attracted her to working at Speechmatics. Jade #### Teri Drummond Director of Engineering https://www.linkedin.com/in/teri-drummond-618a5266/ ### Maria's story Events & Customer Marketing Lead - [Maria's story](https://www.speechmatics.com/careers/meet-our-events-and-customer-marketing-lead) #### Office Manager to Global Brand Builder We sat down with our Events & Customer Marketing Lead to explore the journey behind the scenes — and what it takes to create moments that matter. Discover how she ensures each event is memorable and valuable for attendees and the advice she would give to new starters. Amber #### Maria Anastasiou Events & Customer Marketing Lead Events & Customer Marketing Lead with 7+ years of experience across the SaaS and healthtech industries. https://www.linkedin.com/in/maria-anastasiou-242a3314a/ ## **Thriving together at Speechmatics** Our employee-driven initiatives provide spaces for connection, learning, and personal growth, ensuring that every voice is heard and valued. ### Women at Speechmatics Jade This initiative, established by our female Speechlings, aims to create a strong, supportive community. It provides a space where women can connect, collaborate, and thrive. ### The Journal Club Jade Think of it as a book club - but for research articles! The Journal Club discusses the latest industry findings. One person presents a recent article, followed by an open discussion. ### The Reading Group Amber Our **Reading Group** is all about meaningful discussions centered around fairness, inclusivity, and best practices in coaching, development and collaboration. ### Peer Mentoring Group Orange This evolving initiative offers team members the opportunity to work closely with senior colleagues, gaining invaluable coaching and development experience. ## Quote Left Jade The supportive team and positive culture make every day enjoyable and motivate me to do my best. I truly appreciate how our company has invested in and nurtured my professional and personal development. This commitment is what sets Speechmatics apart from the rest. Jeslyn Yeung Sales Engineer at Speechmatics ### Dan Cochrane quote Left Jade "The team at Speechmatics is fantastic. It's been a dream working with such open, intelligent, and interesting individuals. I've been able to kickstart my career here and grow with the company across different parts of the business." Dan Cochrane Machine Learning Engineer ## Early Careers Internships Amber ### **Launch your career with us** Internships Amber Grow alongside industry leaders Our 12-week internship program is designed to immerse Machine Learning enthusiasts in a real-world environment, including: - **Comprehensive Training** with cutting-edge technology. - **Impactful Projects **contribute to real-world outcomes. - Receive **personalized mentorship** from experienced professionals. - **Team Collaboration** with some of the brightest minds in AI. - [From intern to innovator](https://www.speechmatics.com/careers/eloy-de-jong-story) ## Timeline ### Tirelessly pushing speech technology forward... Throughout the years, Speechmatics has remained at the forefront of speech recognition research and innovation. We are consistently pushing the boundaries of what's possible in speech-to-text technology. #### Timeline - years: - 2015: - Published Research: Scaling Language Models: - description: Our groundbreaking research, as detailed in the paper Scaling Laws for Neural Language Models, explores how the performance of neural language models improves predictably with scale. By analyzing models ranging from small to very large, we established key scaling laws that guide the efficient design and training of advanced language models, leading to significant enhancements in their accuracy and capabilities. - link: [Read the paper in full](https://arxiv.org/pdf/1502.00512) - 2019: - Feature: Neural Punctuation: - description: The first heavyweight, scaled-up punctuation model on the market for use in transcription. This advancement uses sophisticated neural networks to understand context and linguistic nuances, dramatically improving the readability and coherence of automated transcripts. - link: [Learn more about Neural Punctuation](https://www.speechmatics.com/company/articles-and-news/speechmatics-launches-advanced-punctuation-to-transform-the-usability-of-transcripts) - Published Research: Texture bias of CNNs limits few-shot classification performance: - description: Delving into the inherent texture bias in Convolutional Neural Networks (CNNs) and how it impacts their ability to generalize in few-shot learning scenarios. Our research highlights that while CNNs excel at recognizing textures, this bias can limit their performance when classifying new, previously unseen classes, pointing to the need for more robust approaches in few-shot learning. - link: [Read the paper on CNN texture bias](https://arxiv.org/pdf/1910.08519) - 2020: - Parntership: what3words: - description: Our collaboration with what3words marks a significant leap in voice technology - enabling users to enter precise what3words addresses through voice commands. This integration combines our advanced speech recognition capabilities with what3words' innovative geolocation system, offering seamless and accurate navigation solutions. - link: [Our parntership with what3words](https://www.speechmatics.com/company/articles-and-news/what3words-speechmatics-partnered-people-enter-what3words-address-by-voice) - 2021: - SSL Model: Hydra: - description: Our pioneering self-supervised model, Hydra, designed to harness vast amounts of unlabeled data. By leveraging advanced self-supervised learning techniques, Hydra autonomously learns from diverse datasets without the need for manual annotations. - link: [Find out more](https://bold-awards.com/project/hydra-project/) - Feature: Advanced Speaker Diarization: - description: This innovative approach accurately distinguishes and segments multiple speakers in audio recordings, even in complex scenarios with overlapping speech. By leveraging advanced algorithms and large-scale data processing, we have dramatically enhanced the clarity and accuracy of transcribed conversations, making it a game-changer for industries relying on precise audio analytics. - link: [See more about our Speaker Diarization](https://docs.speechmatics.com/features/diarization#speaker-diarization) - Feature: Numerical Entity Formatting: - description: Improvements in transcript readability, for example adding conversions from “four million pounds” to “£4m” and three sixteenths to 3/16. Whether dealing with dates, times, currencies, or other numerical data, this feature intelligently formats and standardizes numerical entities, enhancing the clarity and accuracy of your transcripts. - link: [Find out more](https://docs.speechmatics.com/features/entities) - 2022: - Feature: Language Identification: - description: Our state-of-the-art Language Identification technology sets a new standard in accurately recognizing and distinguishing between multiple languages in audio streams. Using advanced neural networks and deep learning techniques, this feature swiftly and precisely identifies the spoken language, even in challenging environments with mixed languages or dialects. - link: [Detect the prominent language spoken](https://docs.speechmatics.com/features-other/lang-id) - Feature: English Finance Language Pack: - description: Specifically engineered for the finance sector, this advanced language pack leverages cutting-edge speech recognition technology to accurately capture and transcribe industry-specific vocabulary. - link: [Unlocking accurate undestanding of financial terms](https://www.speechmatics.com/company/articles-and-news/speechmatics-unlocks-accurate-understanding-of-financial-terms) - Feature: Language Model Adaptation: - description: Offering a pioneering approach to customizing and refining language models for specific use cases. By dynamically adapting to new data and context, this feature enhances the model's ability to understand and accurately transcribe domain-specific terminology and nuanced language. - link: [See our full list of features](https://www.speechmatics.com/product/features-and-deployments) - 2023: - New ASR Engine: Ursa: - description: Our groundbreaking speech-to-text engine that sets a new benchmark in transcription accuracy. Our next-generation deep learning system trained on millions of hours of audio, is designed to capture spoken words with unparalleled precision, even in noisy or challenging environments. - link: [The world's most accurate speech-to-text system](https://www.speechmatics.com/company/articles-and-news/introducing-ursa-the-worlds-most-accurate-speech-to-text) - Feature: Bilingual English - Spanish: - description: A global approach to language transcription that simplifies the complexity of engaging with multilingual audiences. This feature effortlessly transcribes speech that switches between English and Spanish, ensuring seamless and accurate text output. - link: [Unlocking bilingual transcription](https://www.speechmatics.com/company/articles-and-news/mastering-bilingual-communication-transcribe-english-and-spanish) - Feature: Universal GPU transcription: - description: Our revolutionary approach to deploying all languages on GPU dramatically enhances the speed and efficiency of our speech recognition systems. Whether for real-time transcription, translation, or analytics, running all languages on GPU empowers businesses with faster, more reliable, and scalable language solutions. - link: [Discover more](https://www.speechmatics.com/company/articles-and-news/autoscaling-with-gpu-transcription-models) - Feature: Translation: - description: The most accurate real-time speech translation API, delivering highly accurate translations across a broad spectrum of languages. Whether converting speech-to-text in real-time or pre-recorded, our system ensures that the nuances and meanings are preserved, facilitating seamless and effective cross-language communication. - link: [Breaking down language barriers](https://www.speechmatics.com/company/articles-and-news/our-new-unified-speech-translation-api) - Feature: Summaries: - description: Harnessing the power of LLMs, Speechmatics can extract even more value from your audio data by providing insights and actionable takeaways. By automatically generating clear and accurate summaries, this technology enables users to quickly grasp the essence of lengthy texts or meetings without missing critical details. - link: [Leverage AI-powered Summaries](https://www.speechmatics.com/product/summarization) - Feature: Sentiment: - description: Understand the emotional tone behind the spoken word. Ground-breaking qualitative insights, all built on accurate transcription. - link: [Uncover emotions at scale](https://www.speechmatics.com/product/sentiment) - 2024: - Feature: Audio Events: - description: Improving accessibility by identifying and labelling non-speech sounds in media - helping understand more than speech. This innovation enhances the context and usability of audio data, allowing businesses to gain deeper insights and automate responses to specific audio triggers. - link: [Discover more about Audio Events](https://www.speechmatics.com/product/audio-events) - New Conversational AI: Flow: - description: Flow gives you best-in-class speech technology to ensure every speech interaction is finally as natural as human conversation. - link: [The ultimate API for voice interactions](https://www.speechmatics.com/flow) - New ASR Engine: Ursa 2: - description: 18% reduction in word error rate (WER) across 50 languages compared to our previous Ursa model, with significant gains in Vietnamese (63%), Arabic (48%), and Polish (39%). - link: [Accurate, real-time ASR across 50+ languages](https://www.speechmatics.com/speech-to-text) - 2025: - Feature: Speaker Diarization breakthrough: - description: 48% reduction in speaker identification errors at 1-second latency, and real-time speaker tracking in milliseconds. 31% more accurate speaker labels than our closest competitor. - link: [Discover more about Speaker Diarization](https://www.speechmatics.com/company/articles-and-news/what-is-speaker-diarization-and-why-does-it-matter-in-voice-ai) - Partnership: Ambarella: - description: A collaboration with Ambarella that brings AI-powered natural language interactions to edge applications. By combining Ambarella's edge AI SoCs with Speechmatics' speech technology, users can now experience seamless, natural device interactions; even in environments without the internet. - link: [Our partnership with Ambarella](https://www.speechmatics.com/company/articles-and-news/speechmatics-collaborates-with-ambarella) ## **Ready to join our team?** At Speechmatics, every voice matters - including yours. View our open positions to see what roles we have available. - [Open positions](https://www.speechmatics.com/company/careers/roles)