About us

We are a born-global company. Our roots are in Silicon Valley as we were Google, Facebook AI and Deepmind researchers prior to founding Talka.

We are a team of previous researchers at Google AI, Facebook AI, Deepmind, and other AI research institutions working on conversational understanding that got together to tackle the deep problem of understanding video conversations and improving human communications where it adds a significant business value.

Meet the team

Norberto Guimarães
Founder and CEO
Norberto has 10 years of AI product experience including founding multiple startups. His previous AI company was acquired by Google. He was also the co-chair of Google's AI Research Conference for the ML track.
Wang Ling
Head of ML
Wang Ling is a PhD graduate from Carnegie Mellon University and Instituto Superior Técnico, where he developed the first wave of character-based neural models for Natural Language Processing and Machine Translation. He was one of the first employees to join Deepmind's Natural Language Processing team working on the generation of formal languages and developing better decoding algorithms for sequence generation tasks. He spends his spare time playing table tennis and reading lots of mangas.
José Félix
Head of Operations
Jose, an MSc graduate in engineering and management with a wealth of experience in tech startups and operations consulting, including his recent role as Chief of Staff and business development at an international video telematics company. An avid traveler, he has explored 45+ countries and recently completed an 8-month world tour. Jose perfectly fits Talka's global outlook.
Zac Sun
ML Engineer
Zac has previously been researching on Heterogeneous Cloud Inference Architecture and later joined Intel as Deep Learning Engineer focusing on inference acceleration on x86 architecture and custom silicon. He is a big fan of vintage film photography.
Ray Huang
SW Engineer
Ray has several years of experience as a software developer. He has helped build scalable backend systems that serve millions of users. He is also obtaining master 's degree in computer engineering at UofT as a part-time student.
Hojin Yang
ML Developer
Hojin worked on research in Conversational Recommendation during his MASc at the University of Toronto. He enjoys participating in machine learning competitions; his team won 2nd place in both the 2018 and 2020 ACM RecSys Challenges. Before joining Talka AI, he worked at Layer6 AI as a machine learning intern. Hojin loves to explore Toronto's unique and multicultural neighbourhoods during his spare time.
Ricardo Grade
SW Engineer
Ricardo obtained a MS in Computer Science with distinction at Instituto Superior Técnico. At Talka he develops and improves scalable and reliable applications that offer a good user experience. He values highly the quality and cleanliness of the code he develops. In his spare time, he enjoys watching TV series, playing video games, socializing with friends, and staying up to date with the latest frameworks and technologies to improve his skills as a developer.
João David
ML Engineer
João has a MS from Instituto Superior Técnico specialized in AI and Distributed Systems with a focus on computer vision solutions. During his free time, he enjoys playing the guitar, bouldering, and watching movies.
Miguel Goncalves
ML Engineer
Miguel got an MSc. degree in Computer Science from Instituto Superior Tecnico, where he specialized in Artificial Intelligence and Intelligent Robotics. He previously worked on computer vision solutions using deep learning for traffic control applications. He also has a knack for crafting software that brings autonomous mobile robots to life. Outside of tech, he's all about cars, loves a good game of padel tennis, and can often be found conquering boulders.
Joana Veloso
Human Conversations Team lead
Joana is a Philosophy Major. She leads the 10+ people in-house human annotation team on multimodal communication. Talka allows her to maintain a connection to her interests in human cognition and the marvelous world of language and communication.
Clara Lemos
Human Conversations annotator
Clara is an Electrical Engineer. She is currently pursuing a PostGraduate degree in Marketing Intelligence. She is also part of the 10+ human annotation team. Being a part of the annotation team provides her with a deeper understanding of human communication. Clara's hobbies include reading and learning new languages.
Maria do Carmo
Human Conversations annotator
Mariana is studying psychology. She loves to learn new languages and travel - currently living in Italy to learn and study Italian. She likes to work at Talka because she can watch people's behavior and learn communication skills.
Luiza Oliveira
Human Conversations annotator
Luiza is studying psychology. Being a part of the annotation team provides her with a better understanding of human behavior and communication, which she is very passionate about it. Her favorite hobbies are traveling and meeting new people.
Abid Mohsin
Tech go-to-market advisor
Abid has held various Go-To-Market leadership roles in technology companies, working deeply with Product teams to define and launch features. He was VP Sales for Rev.com, building five sales teams from the ground up. Prior to Rev, Abid managed an Enterprise Sales team at Google, was an early employee managing marketing at Upwork, and spent time in McKinsey. Currently he is helping launch new ideas inside X, Google’s moonshot factory.
Ziyu Wang
Technical Advisor
Ziyu earned his PhD from Oxford in Machine Learning and was one of the early employees of DeepMind working on Reinforcement Learning, Robotics and Bayesian Optimization. He recently hopped across the pond to Google Brain before deciding to pursue entrepeneurial ventures.

What is our ultimate mission and why are we doing this?

Humans are fundamentally social beings.

Talka improves human relationships by improving their communication patterns and mutual understanding. Communication is the bedrock of human relationships. In both business and personal settings.

We strive for impact and push the boundaries of what AI can do to improve that communication by focusing on the deep insights from conversational patterns . With better communication we will have fewer broken relationships both at work and on a personal level and ultimately more fulfilling lives.

Why is this hard?

93% of communication in face-to-face occurs beyond the content of the words. More than the text, it depends on the tone, inflection, body language, and a multitude of other conversational patterns beyond the content. Conversations are very nuanced and the signals are hard to detect. Existing products are only scratching the surface with their basic text and sentiment features.

Our headquarters

We’re headquartered in the San Francisco Bay area in Mountain View, California.
We also have a subsidiary and teams working in Canada, and in Portugal.