Senior Research Scientist at PolyAI | PhD in AI | Building conversational AI that is fast, accessible, and human-centred
Cambridge, England
Joined February 2017
- Tweets374
- Following584
- Followers509
- Likes1.4K
On 4th Nov @matthen2, @pmckee and I are going inside Dialog-RSN-1
When a caller goes quiet, are they finished or just thinking?
Mid-utterance pauses are typically longer than gaps between turns, so getting this wrong interrupts users, or ignores them entirely
Register below 👇
“Hello? Is anyone there?” 🤔
That’s what callers say when a voice agent takes a beat too long. AI can write production code, but many agents still can’t tell when to speak on a live call.
Dialog-RSN-1 hears the caller’s audio and handles the whole turn in a single pass, so callers hear a reply in under 1 second. Independent benchmarking by Coval found it the strongest realtime model they’ve tested on production-style calls.
@pmckee, @Addlesee_AI, and @matthen2 take you inside the model in Part 1 of Let’s Dialog: Building the Next Voice Agents.
Grab your spot: bit.ly/4dVUc6B
I first wrote about mid-utterance interruptions due to pauses 7 years ago!
Today we launched Dialog-RSN-1, and it pauses with you. If you say "my ref ID is 4, 7, 1…" and go quiet while you check the next digits, it waits.
See the full video demo here 🎉
poly.ai/dialog-rsn-1
Dialog-RSN-1 is now GA, and anyone can get an API key and build on it today.
Comparing real-time voice models? Start with the numbers. Median first-token latency is under 300ms, about the length of a blink. It has the highest quality, audio-understanding, and safety scores of any real-time-capable model, and it’s efficient enough to scale to every conversation you have.
It’s audio-native, so turn-taking, speech recognition, function calling, and the response all happen in a single pass, with a separate voice output you control.
Get started in three steps:
1. Pick your path: our Agent Development Kit (ADK), AWS SageMaker, or the public API, wherever your stack already lives.
2. Create your account and grab your API key.
3. Clone an example, add your key, and talk to it in your browser.
Start now here: studio.poly.ai/start-buildin…
Apply here: amazon.jobs/en/jobs/3126764/…
And if you haven't tried Alexa+ yet, sign up here: amazon.co.uk/newalexa
We're looking for interns to join us at @AmazonScience! Come build the future of AI with us.
We just launched Alexa+ in the UK and there's so much more to come 🎉
MS/PhD students in GenAI, LLMs, speech, NLP & more - this one's for you. Based in London or Cambridge.
Links in 🧵
Angus Addlesee retweeted
📢 Call for Papers: #SIGDIAL2026
Submit your latest work on discourse & dialogue to the 27th Annual SIGdial Conference (Atlanta, July 13–15, 2026)!
🗓 Deadlines (AoE)
• Mar 23 – Title & abstract
• Mar 30 – Full paper (PDF)
• May 4 – ARR commitment
• May 25 – Notifications
This was a major collaboration across Europe, huge thanks to:
@oliverlemon for leading the team from @HeriotWattUni, and all SPRING partners & colleagues who made it possible (see the huge author list!)
A milestone for socially assistive robotics in healthcare! 🤖💬
The paper ties together work from 8 institutions to advance human–robot interaction (HRI) in real healthcare settings
We address:
• Robot navigation in clinical environments
• Speech processing & non-verbal behaviour
• Accessible conversational AI for older adults
• & more!
Our SPRING Project paper has just been published in the International Journal of Social Robotics! 🎉
📄 Socially Pertinent Robots in Gerontological Healthcare
We explore how hospital patients & our robot Ari can engage in natural, useful conversations
link.springer.com/article/10…
It has been wonderful collaborating with my friends at @HeriotWattUni 's Interaction Lab (@iLab_hwu). Thanks!
In January, I submitted my PhD thesis titled: "Incremental Multi-party Conversational AI for People with Dementia"
Luckily, even the sun came out for graduation day ☀️
Congratulations to artificial intelligence graduate John-Angus Addlesee, who is the third generation of his family following his father and grandfather to study at Heriot-Watt University. 🎓👏
#HeriotWattUni #HWUGrads
Angus Addlesee retweeted
Thrilled to get this paper out, a great collaboration with @UCL_VoCoLab, @Addlesee_AI and @verena_rieser on how self-bias and sense of agency over a voice is altered when that voice is used to represent oneself in a social interaction.
Self-ownership, not self-production, modulates bias and agency over a synthesised voice
📢New from: Bryony Payne, Angus Addlesee, Verena Rieser, & Carolyn McGettigan
sciencedirect.com/science/ar…
Angus Addlesee retweeted
Multi-party Multimodal Conversations Between Patients, Their Companions, and a Social Robot in a Hospital Memory Clinic
Angus Addlesee, N. Cherakara, N. Nelson, D.H. Garcia, N. Gunson, W. Sieińska, C. Dondrup, O. Lemon
aclanthology.org/2024.eacl-d…
Best Demo Award #eacl2024
I am delighted to announce that we won the "Best Demo Award" at #EACL2024 🎉
It can be found here: aclanthology.org/2024.eacl-d…
Thanks @eaclmeeting!
Authors: @Addlesee_AI, @n_cherakara, Nivan Nelson, @d_hgarcia, @nagunson, Weronika Sieińska, @cdondrup and @oliverlemon
@iLab_hwu
