Food Standards Agency

Strategic Surveillance delivery

8 Incomplete applications

7 SME, 1 large

13 Completed applications

6 SME, 7 large

Important dates

Published
Wednesday 26 February 2020
Deadline for asking questions
Wednesday 4 March 2020 at 11:59pm GMT
Closing date for applications
Wednesday 11 March 2020 at 11:59pm GMT

Overview

Summary of the work
Ensure we build on the success in delivering on the pipeline of use cases, to continue innovating and improve further our situational awareness for food and feed risk the Food Standards Agency requires a supplier to augment our internal team with highly skilled multi-disciplinary individuals/teams
Latest start date
Monday 20 April 2020
Expected contract length
2 years
Location
London
Organisation the work is for
Food Standards Agency
Budget range
The total value of the proposed call-off agreement will not exceed £3M.
Work packages will be raised under this proposed call-off approach for each project.
Work packages will be on a capped time and material payment profile.
The FSA will not guarantee any minimum value of work under the proposed call-off agreement.

About the work

Why the work is being done
Our goal over the last 24 months has been to extend upon the systematic operations across the FSA to effectively identify food and feed risks. We need to identify direct impacts such as a microbiological risk through to crime, changes in consumer attitude to food to changes in business models, then to technical advancements in the production of food and how we all engage with it. We must be able to protect consumers now, and in the future, with the foresight to predict and take action in a timely manner.
Problem to be solved
To optimise use of data, mainly open data, to identify emerging food and feed risks before they become a public health concern.
Who the users are and what they need to do
As the competent authority in charge of regulating the food and feed sector, the FSA needs to be aware of risk affecting UK consumers, whether related to food and feed safety or to authenticity. Developing a predictive capability within the Agency will allows us, and others, to take action to mitigate those risks.
Early market engagement
Any work that’s already been done
Please see Document linked below for an explanation of previous work and additional information on requirement:

https://drive.google.com/file/d/1Kkh4FDSC3vhJDQwPYkqQXSMN--XtxcOU/view?usp=sharing
Existing team
The Supplier will be working with an in house team consisting of:
1 * Head of Data
3 * Data Scientists
1 * Fast Streamer (data analytics experience)
2 * Research Analysts (part-time)
Current phase
Live

Work setup

Address where the work will take place
Food Standards Agency, Floors 6 + 7, Clive House, 70 Petty France, Westminster, London, SW1H 9EX
Working arrangements
You will be expected to work openly and collaboratively on digital channels (eg slack, MS Teams, Office365 etc). You will be expected to attend face-to-face key meetings at FSA offices, be involved in catch-up calls, organise and run user testing sessions, and present findings.

We expect the successful supplier to provide upskilling to permanent staff to increase internal capability.
Security clearance
All supplier team members will need to be cleared to Baseline Personnel Security Standard, and may be requested to sign a non-disclosure agreement (NDA)

Additional information

Additional terms and conditions

Skills and experience

Buyers will use the essential and nice-to-have skills and experience to help them evaluate suppliers’ technical competence.

Essential skills and experience
  • Development of ML models and prototype dashboard for intuitive predictive insights
  • Strong technical knowhow of various Machine Learning platforms, tools, and algorithms
  • Following an agile way of working centred around specific ‘use cases’ and delivering Prototypes / Minimal Viable Products
  • Experience of data cataloguing, utilising Github for version controlling and deployment of dashboards
  • Experience of building machine learning models is multiple programming languages: Python, R, SQL
  • Experience of using Microsoft Azure platform for developing end to end pipelines for running AI / ML models: data ingestion, processing, modelling, and visualisation
  • Experience of visualising AI / ML based outputs using R Shiny and / or Power BI
  • Experience of working in multiple Data Science Libraries such as TensorFlow, Pytorch, Keras, Pandas, SKLearnSpacy, beautifulsoup, Numpy etc.
  • Knowledge of conducting detailed analysis of datasets available – ingesting, cleaning, analysing and modelling data as per business needs using best practices
  • Liaising with different parts of the business and with external entities to support the forward pipeline of use cases and areas for further collaboration
  • Capturing use case requirements, other technical documentation related to the use cases in a systematic format
  • Development of the analytical roadmap, implementation plan for operationalisation of use cases
  • Experience of delivering business value in similar context – making sure business requirements are understood, validated via positive challenge and delivering value
  • Demonstrable deep knowledge and applied experience of the Government Data Ethics Framework in AI/ ML models
  • Demonstrable approach to ensure reusability by design – ensuring solutions are formulated from reusable capabilities as well as ensuring new capabilities are reusable
Nice-to-have skills and experience
  • Familiarity/understanding of food & feed related risk and how data science may be applied to related risk signals
  • Experience of collaboration software such Jira, Confluence, Trello, Slack and Teams
  • Experience organising hackathons / stakeholder workshops to agree business requirements and necessary solution approach
  • Experience working with public sector organisations
  • Experience of how to develop, code, test, correct and document programs as part of a multidisciplinary team

How suppliers will be evaluated

All suppliers will be asked to provide a written proposal.

How many suppliers to evaluate
3
Proposal criteria
  • Demonstratable delivery of solutions providing insight and delivery of measurable outcomes using AI and predictive modelling, including machine learning, neural networks and deep learning with disparate api/linked open data sources.
  • Deep knowledge and applied experience of the Data Ethics Framework and supporting guidance from The Alan Turing Institute for the responsible design and implementation of AI systems in public sector.
  • Demonstratable experience working with and leveraging the benefits of Microsoft Azure services and open source packages for efficient and robust end-to-end solutions.
  • Experience of end-to-end lifecycle from discovery to operationalising production grade services, ensuring quality assurance, adherence with relevant legislation, code of practice regulation and guidance related to use of AI/service design.
  • Your approach to delivering business value – making sure business requirements are understood, validated via positive challenge and delivered against while understanding the value being generated (i.e. business intent)
  • Evidence of innovative thinking - how you promote thinking outside the set of existing organisational constraints to arrive the most appropriate and efficient route to deliver desired outcomes
  • Demonstratable approach to AI governance to ensure models and their delivery process can generate results which facilitate explainable AI by being reproducible, traceable and verifiable.
  • How you will ensure a methodical and measurable approach to knowledge sharing to support our development of in-house skills and experience and mitigate supplier dependencies
  • Demonstrable approach to ensure reusability by design – ensuring solutions are formulated from reusable capabilities as well as ensuring new capabilities are reusable
  • Demonstrable approach to simplicity by design with a clear focus on producing and positively challenging solutions to be as simple as possible
  • Demonstrable approach to identifying risks, dependencies and proposing mitigations throughout the lifecycle
  • Proposed team structure, capabilities and experience
  • Enough depth and breadth of capability across the supplier resource pool and ability to flex as needed
  • Describe how your proposal will optimise costs and state any additional value-add your organisation will provide
  • Value for money
Cultural fit criteria
  • Be transparent and collaborative when making decisions
  • Experience of successful collaborative working as part of a multi-supplier agile delivery team, sharing knowledge and experience within the team
  • Able to challenge the status quo and use persuasion skills to achieve business buy-in
  • Take responsibility for their work and a no-blame culture and encourage people to learn from their mistakes
  • Bring agile mindset and principles into a developing organisation
Payment approach
Capped time and materials
Additional assessment methods
  • Case study
  • Presentation
Evaluation weighting

Technical competence

60%

Cultural fit

15%

Price

25%

Questions asked by suppliers

1. Is there an existing incumbent for this project?
Yes. The incumbent supplier is Cognizant.
2. Will all the work have to be carried out in your office in London with travel to other sites or can the work be done remotely from supplier location with travel to your offices as need arises?
We are open to discussions around the format for delivering the work but envisage most will be conducted on FSA’s premises in London.

We have found that a face to face relationship between the data science provider and the business experts (internal FSA colleagues) is essential.
3. Please can you let us know which external suppliers have completed the previous phases of work?
Cognizant are the incumbent supplier and have completed previous phases of work.
4. In the linked Strategic Surveillance – 2020 ITT document, it states the in-house team is augmented with the supplier team (1 x Senior Manager, 3 x Senior Data Scientists, 3 x Junior Data Scientists etc). Is this an incumbent supplier working with the FSA or is it just a description of the roles you would anticipate in the supplier team to deliver this project? If an incumbent supplier, are you able to say who the supplier is and if they will be bidding?
The roles listed (1 x Senior Manager, 3 x Senior Data Scientists, 3 x Junior Data Scientists etc) are the details of the incumbent suppliers team.

The Incumbent is eligible to bid for this requirement, however we are not aware if they intend to.
5. Please confirm if you will require all roles stated for this requirement to start from day 1
All roles should start on day one.
6. Can you please confirm if you have in place a Delivery Manager / Project Lead/ Testers for your requirement, or would the award supplier be expected to provide these candidates?
Strategic Surveillance has an FSA service lead. The service lead gives strategic shape and direction, as well as day to day management, of the work of the Strategic Surveillance team.
7. Do you want the incumbent supplier to bid for this opportunity, and have you or anyone at the FSA encouraged them to do so (even informally)?
The Incumbent supplier is aware that the current contract is coming to an end and that the FSA are in the process of tendering for a new supplier.

To ensure impartiality the FSA has not discussed the new tender with the incumbent.
8. Will you accept joint bids (from two suppliers, both on the DOS framework), or sub-contracting arrangements (again, where all suppliers are on DOS)?
The FSA will accept joint bids and/or bids containing sub-contracting arrangements for this work.

In the case of joint bids, the FSA would look to have the contract in the name of one supplier only.
9. Please can you share a procurement timeline outlining steps (which would typically include evidence response evaluation, shortlisting of suppliers for next stage, preparation of supplier proposals, evaluation of supplier proposals, announcement of successful supplier, contract completion, contract signatures)?
The estimated timeline is below. These may vary, dependant on how many applications are received. The intention is for the contract start date to be 20/04/2020.

26/2/2020 – 11/03/2020 – Opportunity open on Digital Marketplace.
12/03/2020 – 18/03/2020 – review of applications and shortlist of suppliers for next stage.
19/03/2020 –invite shortlisted suppliers to assessment stage and provide feedback to unsuccessful suppliers.
19/03/2020 – 02/04/2020 – shortlisted suppliers proposal preparation, submission date of 02/04/2020.
03/04/2020 – 13/04/2020 – Evaluation of proposals.
13/04/2020 – 17/04/2020 – Announcement of successful supplier, contract preparation and signature.
20/04/2020 – Contract Start date.
10. Would non-public sector case studies be treated on an equal footing as public sector ones?
Yes, both public sector and non-public sector case studies will be treated on an equal footing.
11. What team roles will be provided from FSA for this engagement?
Team roles provided by the FSA for this engagement will be:

1 * Head of Data
3 * Data Scientists
1 * Fast Streamer (data analytics experience)
2 * Research Analysts (part-time)
12. Which of your stated use cases are currently live?
The use cases that are currently live are below:
Risk Likelihood Dashboard
Meat Establishment Dashboard
Aflatoxin risk prediction
Signal prioritisation
Pesticide risk prediction
Food Consumer experience
Trade Routes and Volumes at Ports
Non-UK RASFFs
Online display of FHRS scores (this is currently live for Wales only)
13. In order to compare suppliers, like for like, how will price be evaluated: will you be evaluating by the total price, rate card or both? How do you ensure that you are completing a genuine value for money comparison rather than just a rate card comparison?
The FSA will be looking to evaluate shortlisted suppliers by asking them to provide the cost for resourcing a 10 week sprint, based on the rates for the staff roles they would look to provide the FSA.
14. Can you provide an example of capability you would want to reuse?
The FSA’s would look to reuse capability whenever we can. For Example, if we were to use weather data for a use case the approach would be to reuse this knowledge whenever we can. This would also apply to any techniques and natural language processes, Bayesian networks, random forests, etc.
15. Can you provide an example of capability you would want to reuse?
The FSA’s would look to reuse capability whenever we can. For Example, if we were to use weather data for a use case the approach would be to reuse this knowledge whenever we can. This would also apply to any techniques and natural language processes, Bayesian networks, random forests, etc.