Join our movement to provide parents with the best possible care
Research Scientist in Conversational AI
The transition into parenthood is tough. Parents face pressures to be a ‘good’ parent; negotiate their identity at work; and raise children with less support than ever before. 1 in 3 parents are experiencing perinatal mental ill health.
Kami’s vision is to empower parents with a personalised, data-driven, evidence-based coach. We are using the insights and experience of all humanity to support individuals on their personal journey into parenthood. The value that we deliver to our users comes from the wealth of knowledge from our experts and coaches, whose intuition and experience allow them to customise the support on-the-fly based on what they pick up from the user. We envision machine learning can short-cut some of these.
We are delivering our support through a judgment-free, artificially intelligent, digital companion that enables parents to:
Get trusted, non-judgmental advice and answers to their questions;
Set goals and monitor progress on their wellbeing and parenting;
Accomplish their goals through a personalised concierge of activities and specialist support;
Access tele-consultation sessions with vetted consultants specialising in fertility, parenting, wellbeing, and behaviour change;
Find and access products and local services curated to their needs.
Applications are invited for a R&D engineer in Machine Learning / Natural Language Processing, with a specific focus on conversational AI (spoken and/or text-based dialog systems) to join us to work on the frontier of Kami’s core technologies. The position offers the opportunity to have a major impact at an early-stage startup. The company is made up of experts from Imperial College, King’s College London and the University of Surrey and you will be part of a very strong team.
The purpose of this project is to develop a platform and chatbot connected to a collection of controlled flow interactions created by therapists and psychologists, a database of certified information and resources, and a panel of experts in various aspects of pregnancy & childbirth (i.e. midwives, doula, parenting coaches, etc).
It is then to enhance the performance of the platform, by creating predictive algorithms to determine the best way to support our users, in particularly preemptively addressing issues and challenges the users might face throughout pregnancy.
Kami challenges you to create solutions that improve the pregnancy and childbirth journey. A significant dataset of over 100,000 online conversations between new and veteran parents -- containing text, images, and video content provided by 1000+ unique accounts -- will be available to you. You will be working with time-series datasets consisting of text data to establish a causal link between temporal and behavioral factors during pregnancy to the likelihood certain (particularly mental health and wellness related) problems or difficulties arising. Clustering the resulting user data into notable groups in their likelihood to have certain challenges and predicting likely profiles of new users. For example, the largest dataset contains 60K rows collected from pregnancy forums; key topics have been extracted but not analyzed to understand correlation to user profile. From this dataset we can identify: unique usernames, title of posts, time of post, #of responses, keywords, tags, geographic location of users. For the majority of the users, we can also identify their time of pregnancy and date of birth.
The role is for a project lead, based in King’s Entrepreneurship Institute, Bush House, The Strand, London, reporting to Helen Yannakoudakis, Assistant Professor at King’s College London in the area of Machine Learning for NLP and a stakeholder at Kami.
Research Scientist (fixed-term contract):
Salary: £19.34 per hour; possibility of equity share options.
Working 21 hours per week, fixed-term contract for 16 weeks; possibility of extending it further.
Flexible location / working considered.
The position offers the opportunity to have a major impact at an early-stage startup.
Appointments will be from 15th July or as soon as possible thereafter.
● Data preparation and processing for training and testing conversational AI systems;
● Research and evaluate state-of-the-art machine learning algorithms in conversational AI;
● Develop, prototype and train algorithms to be integrated on the platform;
● Deliver software prototypes for deployment, ensuring the highest standards of quality and respect for user privacy.
● A PhD in Machine Learning / NLP with a focus on conversational AI.
● Experience with Machine Learning, Deep Learning and NLP frameworks (e.g., TensorFlow).
● Strong command of at least one of the following production programming languages C# / C++ / Clojure / Python / Java.
● Practical knowledge and experience in building algorithms and designing experiments to merge, manage, interrogate and extract data.
● Self starter with strong independent problem solving skills, great communication skills and
the ability to collaborate with a team.
● Practical experience with both spoken and text-based dialog systems.
● Practical experience in developing multi-modal machine learning models.
● Experience in developing machine learning solutions under deployment constraints such as limited compute resources.
● Being able to integrate the ML/AI within a commercial platform.
● Having developed a ML module for a deployed iphone/android app.
● Experience with content-based data and image retrieval, visual search.
● Polyglot programmer.
● Database build and management in a professional environment.
● Desire to be part of affecting real change and create a product with a positive impact in our society
● Driven self-starter, fast and keen learner
● Being curious and interested in solving societal problems
To apply forward your CV with a full list of publications and a cover letter (2-3 pages) including a summary of your research and product development experience, as well as how you meet the selection criteria for the post. You should also provide contact details of three referees.