This opportunity is closed for applications

The deadline was Monday 31 October 2022
Department for International Trade

Department for International Trade - Data Science Innovation Lab

8 Incomplete applications

5 SME, 3 large

28 Completed applications

15 SME, 13 large

Important dates

Published
Monday 17 October 2022
Deadline for asking questions
Monday 24 October 2022 at 11:59pm GMT
Closing date for applications
Monday 31 October 2022 at 11:59pm GMT

Overview

Off-payroll (IR35) determination
Contracted out service: the off-payroll rules do not apply
Summary of the work
A Data Science team is required to work on complex data science problems to deliver solutions that help DIT to meet its objectives. Each solution is expected to take between 2 weeks to 3 months.
Latest start date
Monday 6 February 2023
Expected contract length
24 months + an optional extension of up to 6 months
Location
London
Organisation the work is for
Department for International Trade
Budget range
Total contract value of £1,750,000 - £2,000,000 (excluding VAT) for the initial 24 month period. DIT reserves the right to extend the contract for a period of up to six months with an additional value of up to £500,000 (excluding VAT).

Each Statement Of Work (SOW) will have a budget allocated to it and the cost of each SOW must be agreed by DIT in writing before the work commences.

About the work

Why the work is being done
The Department for International Trade (DIT) requires a specialised data science team to deliver innovation projects drawing upon data science tools and techniques to meet DIT objectives.

There is an existing Data Science function with DIT. The existing teams provide support to existing business units, however there are some projects that need further development and cannot be supported by the current teams.
Problem to be solved
There are a range of problems to be solved as part of this contract, including:

- Applying Data Science to improve services DIT provides to traders and the support we give to DIT staff. For example, using data science to improve how we generate leads, understand natural language text and deal with enquiries.

Developing tools and features as part of wider products that enable DIT to meet its objectives around:

- Delivering economic growth to all the nations and regions of the UK through attracting and retaining inward investment.

- Supporting UK business to take full advantage of trade opportunities, including those arising from delivering trade agreements, facilitating UK exports.

- Enabling the specifics of getting goods and services across a border and into a market.

Each project will be set out in a Statements of Work with it owns specified outcomes. Statements of work will be agreed over the lifetime of the Contract.
Who the users are and what they need to do
The users are DIT's internal staff and external users who access public facing digital services.
Early market engagement
Not undertaken
Any work that’s already been done
Existing team
There is an existing data science function within the DDaT directorate in the Department. Data Scientists within DDaT are embedded into product teams, using data science to inform product development and deliver data science features.

The appointed supplier would be expected to provide full teams and / or individuals for the purpose of delivering specific outcomes within existing DIT teams. When full teams are provided, the supplier may be required to provide additional roles, such as delivery manager, in addition to the core data science capability. The exact requirement will depend on business needs.
Current phase
Live

Work setup

Address where the work will take place
Old Admiralty Building, London, SW1A 2BL
Working arrangements
The team operates under hybrid working arrangements. DDaT has staff in various locations in the UK - the main locations are London and Cardiff. Therefore, exact working arrangements will be determined with the successful supplier.
Security clearance
All supplier staff will be required to have SC clearance before they start. A copy of the clearance from the supplier will be required. It is the responsibility of the Buyer's contract manager to ensure clearance is received.

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
  • Experience using natural language processing and core machine learning techniques. (2%).
  • Expertise in agile ways of working in delivery of data science projects. (2%).
  • Thought leadership and knowledge transfer in the use of machine learning. (2%).
  • Experience telling engaging stories with data by using effective data visualisation tools and techniques. (2%).
  • Expertise around deep learning & neural networks. (2%).
  • Experience working with highly unstructured data and make it fit for use in data science projects. (2%).
  • Experience delivering as part of a wider product team, under the leadership of a programme team, as part of a wider portfolio. (2%).
  • Demonstrate the ability to work with stakeholders to refine and validate their ideas. (2%).
  • Strong culture and experience around best practice in software engineering as applied to data science. (2%).
Nice-to-have skills and experience
  • Demonstrate experience in the planning and delivery of user-centred, agile projects, following the Government Service Standards and Technology Code of Practice, ideally via demonstration of meeting a service assessment. (1%)
  • Demonstrate experience working as a strategic supplier for another organisation to deliver long-term data science projects. (1%)
  • Demonstrate international trade (or related field) domain expertise. (1%)
  • Demonstrate experience working with the AWS platform. (1%).

How suppliers will be evaluated

All suppliers will be asked to provide a written proposal.

How many suppliers to evaluate
3
Proposal criteria
  • Demonstration of real innovation in the application of data science techniques to solve real world problems, i.e. going beyond standard tools and technologies. And how this will assist you-to-deliver-this-requirement-to-DIT,-explain-how-you-will-apply-these-experiences-to-DIT? (6%).
  • Describe the diversity of individuals in your team set-up for this contract (including roles and experience), explain the processes in place to decide what goes forward as part of-discovery-work-and-how-you-would-adapt-the-team-to-changing-priorities. (4%).
  • Describe how you will ensure knowledge transfer to DIT staff will take place after each project. (4%).
  • Describe how you will use data science to help deliver innovation in DIT, and how this reflects the unique circumstances of DIT. (3%).
  • Identify any risks and dependencies that you foresee for this project, and offer approaches to manage them. (3%).
  • Detail what your potential Discovery report would look like for projects, and how you will ensure the outputs can be used for any potential Alpha phase. (3%).
  • Detail how you will ensure user needs are understood and met in the development and delivery of the service. (3%).
  • Please demonstrate how you will introduce novel data science solutions, to help DIT solve problems in innovative ways to support their existing data science capability. Please-include-reference-to-3- engagements, include-one-where-you-operated-independently-and-one-as-part-of-an-inhouse-team.(10%).
  • SOCIAL VALUE: Theme 4: Equal Opportunity. Policy Outcome: Tackle workforce inequality. Criteria 6.2. Describe the commitment your organisation will make to ensure that opportunities under the contract-deliver-the-policy-outcome-and-award-criteria.(5%).
  • SOCIAL VALUE: Theme 5: Well-being. Policy Outcome: Improve Health and Well-being. Criteria. 7.1. Describe the commitment your organisation will make to ensure that opportunities-under the-contract-deliver-the-policy-outcome-and-award-criteria.(5%).
Cultural fit criteria
  • Demonstrate your ability to integrate and work alongside existing Data Science & development teams. (3%).
  • Demonstrate how you will communicate effectively across all levels of DIT. (3%).
  • Demonstrate your ability to deliver in an open, collaborate, agile way according to the principles outlined in the Government Service Design Manual. (3%)
  • Demonstrate how you will build Data Science skills and capability within the existing data science team. (3%).
Payment approach
Capped time and materials
Additional assessment methods
Presentation
Evaluation weighting

Technical competence

68%

Cultural fit

12%

Price

20%

Questions asked by suppliers

1. The SC clearance restriction is firmer than we see at many MoD projects and it would likely rule out a lot of SMEs from competing. I'm just checking this will be necessary for all DIT projects and no sponsorship can be provided by DIT? If it's not 100% of the time but will be needed, has DIT considered making 2 lots - 1 for SC cleared staff and 1 for non-SC cleared staff? This would give DIT more choice and flexibility.
SC is required for all data scientists working with DIT data and is needed for writing code and bulk access to data. So this will be a necessary requirement for data scientists. Potentially other roles without build access to data such as delivery manager or other oversight roles could be involved without SC.
2. For the nice to have skill: Demonstrate international trade (or related field) domain expertise. What do you consider to be a related field?
Related fields may include (but not limited to):
- logistics
- supply chains
- investment and lead generation
- regional UK business operations
- general collaboration with UK government organisations.
3. The second competence in 'Nice to have' states: 'Demonstrate experience working as a strategic supplier for another organisation to deliver long-term data science projects.' Can you please confirm whether 'another organisation' means seperate to DIT (and so DIT cannot be the example used)?
'"Another organisation" may include DIT.
4. Can the authority please confirm the incumbent providing this service.
The previous contract has concluded. The previous supplier was PA Consulting Services Ltd.
5. What are the programming languages being used for AI/ML model development and can you please confirm the full technology stack that is in scope?
Python-is-the-core-language-being-used-for-development-of-AI/ML-models. Almost-all-of-the-current-codebase-is-in-Python. SQL-is-also-used-for-data-extraction/transformation.
In-terms-of-the-technology-stack, the-department-has-a-collection-of-cloud-based-tools-available-on-our-data-workspace.
This-includes:
- visualisation-tools (such as Quicksight)
- analysis-tools (such as pgAdmin)
- development-tools (Jupyterlab Python, Rstudio and Theia)
- data-management-tools (Gitlab)

DIT's data-workspace-is 'insulated' from-the-internet-and-not-every-feature-or-package-is-available-by-default-but-most-requirements-can-be-onboarded-in-collaboration-with-our-data-infrastructure-team. (Access to external data sources on the internet is carried out via AirFlow).

The data science team is constantly exploring new tools that could enhance data science development and deployment. For example, the team currently uses MLflow as well as several key machine learning libraries. Suggestions for new tools or features are encouraged. We are looking for the successful supplier to help inform our choices in terms of technology stack.
6. Are you able to provide us the existing data governance and/or security requirements (protocol) for DIT data science projects?
All internal data within DIT has a named information asset manager, who oversees and manages access to each dataset, in line with the legislation and government security classification procedures. All data science work will be carried out on our data platform. New tools, applications and uses of data must go through our own internal accreditation process, which will include a penetration test if necessary and oversight from our Data Protection team and relevant departmental leads.
7. Can you please provide us with a summary of use-cases that have been completed in the last 2 years and success/impact they made?
FTA use case:
Extraction-and-summarisation-of-FTAs-to-enable-and-aid-content-development.
This-use-case-involved-extracting-text-from-free-trade-agreements-into-a-structured-format. The-resulting-data-structure-can-be-queried-to-output-relevant-extracts-and-then-those-extracts-can-be-summarised. The-purpose-of-this-is-to-surface-key-information-from-FTAs-around-a-certain-topic-or-theme-in-a-quick-and-accessible-manner-as-FTAs-can-be-long-and-difficult-to-read. The-summarised-extract-can-then-be-reviewed-and-edited-by-content-editors-for-potential-publication-on-DIT's-various-online-services. This-use-case-is-still-being-under-development, but-is-being-used-by-our-Content-team.

Tariff-checking
This-use-case-involves-automated-extraction-and-comparison-of-tariff-data-sources. The-UK's-tariff-is-defined-in-various-legislative-documents-and-is-stored-in-DITs-data-systems. Updates-may-be-made-at-different-times-which-can-result-in-discrepancies developing. This can have negative repercussions as rates that traders are charged at customs may be different to that defined in legislation. The tariff checking tool extracts data from various legislative documents, compares this to data stored in systems and highlights any discrepancies. The algorithm developed attempts to account for false positives and uses various techniques to make comparisons between duty rates, commodity codes and commodity descriptions. This tool is used periodically, prior to updates to legislation, in order to enhance confidence in the UK's tariff.
8. Will any of the completed use-cases need further enrichment if yes can you provide some details?
A number of proof-of-concepts have been delivered and so yes they will likely need further enrichment. Many of the use cases use natural language processing techniques.
Current work underway by the data science team includes which could benefit from enrichment includes:
-Company matching
-Lead generation
-NLP to support content and Enquiries.
9. Can the authority provide details of the key challenges in the current programme?
Key challenges include:

- High quality documentation and effective handovers to enable knowledge transfer so that work can be continued easily
- Thought leadership in the Data Science space
- Understanding the DDaT context and the service standard in order to develop high potential areas for application of data science.
- Making outputs engaging for non technical audiences in order to build relationships, maintain focus and deliver new opportunities
- Maintaining progress in an environment of shifting priorities, with a focus on continual release.
10. Is there any part of the current data science/ML lifecycle that is proving most challenging (e.g., feature engineering, model monitoring)?
The current ML model deployment and monitoring process can be difficult. The team is in the process of changing out workflow in this area, which should make deploying retrained models easier as well as monitoring their performance in a structured way.

We have some limits on our platform around available memory so in some instances very large datasets or even models can present an issue (although this is an uncommon issue). We are also currently working on solutions to this problem.
11. Can the authority share the current team size, structure, and role of the team provided by the current incumbent and DIT inhouse team?
The-DIT-inhouse-team-is-currently-made-up-of:
The-head-of-Data-Science, with-5-senior-data-scientists-supported-by-3-Data-Scientists-and-3-Junior-data-scientists. The-in-house-data-science-team-are-allocated-to-various-portfolios-within-the-department-that-focus-on-different-work-areas-such-export-and-investment-or-technology. Most-of-the-in-house-data-scientist-time-is-split-across-different-projects. There-are-separate-teams-for-Data-Analysis, Data-Engineering, and-Data-Architecture-as-well-as-the-wider-DDaT-professions. All-roles-work-closely-together-in-multi-disciplinary-agile-teams.

The-previous-supplier-team-was-made-up-of-4-to-5-individuals-on-average. They-focused-on-specific-data-science-problems-and-would-provide-a-dedicated-team-of-individuals-to-rapidly-develop-a-proof-of-concept. They-were-supported-by-the-civil-servants Data Scientists in engaging with stakeholders. An civil servant data scientist would also provide oversight to ensure that outputs are acceptable and to enable continuity once a project was delivered.
However it is expected than the focus going forward will be on highly skilled data science specialists who can set the direction in terms of application ongoing data science and new projects. Since the Data Science team has now been built up we do not need have as much need for standard or intermediate level data science resource.
12. Has a period for transition being agreed with the incumbent and what will be the overlap period?
There will be no transition period as the previous team have fulfilled their contract and are no longer working with DIT.
13. Will the authority consider an innovative proprietary AI-driven tool that could be considered as going beyond standard tools and technologies (listed on GCloud) if it assists in the delivery of this requirement?
The-data-science-team-welcomes suggestions for new tools and technologies that can enhance delivery of their work. The general preference amongst the team is for open source tools. If a propriety tool were to be considered it would need a robust use case.
Typically we only use Open Source software in line with the service standard. We would not rule out using proprietary software if the was a clear use case. However we would likely have to look in detail at the product to understand if it can be used on our Data platform and is consistent with our departmental ways of working.
14. Can the authority please provide additional context to the question below? 'Strong culture and experience around best practice in software engineering as applied to data science.' By Software Engineering do you mean AI/ML engineering?
This refers to conforming to industry standards around code development. Code should be well written and documented. For python, for example, the expectation would be to conform to the pep8.
15. Can the authority provide some details on the status or maturity level of the data science platform currently in operation?
The Data Science platform in use is a live service in Public Beta (all users are internal to government) and is the Data Platform for the whole department, i.e. it also provides dashboarding capabilities for all staff and allows downloading of raw data among a range other features. The service is undergoing continuous improvement as part of business as usual work.
16. Can you provide us details on the current end-to-end process for new use cases from discovery to onboarding or operationalisation?
In general, new-projects-will-try-to-follow-the-GDS-service-standard-and-work-within-an-agile-delivery-framework. Please-see-the-links-for-further-information:
https://www.gov.uk/service-manual/service-standard
https://www.gov.uk/service-manual/agile-delivery

For-data-science-specific-problems, the-workflow-may-sit-somewhat-outside-of-the-process-mentioned-above. Many-data-science-projects-may-follow-the-following-process:

- Problem-identified-that-could-be-solved-with-data-science. This-may-be-something-data-scientists-identify-themselves-or-via-product/service-owners
- Explore-available-data-that could solve problem
- agree and confirm that data science should/could tackle this problem
- meet with stakeholders to develop requirements
- develop data science solution proof of concept. This may include some basic visualisation
- agree and identify opportunities for further development
- refine solution
- deploy model or solution into production. This could mean deploying a model into production so that its predictions are usable or could mean engaging with other professions such as developers/designers to develop or integrate the proof-of-concept into a product.
17. Will this new data science team be required to work end-to-end from discovery to operationalisation and ongoing improvement or focus purely on proof-of-concepts for use cases/model development?
The requirements of the department will ultimately dictate which stages the data science team will be required to work on. Previously, the focus has been on developing proof-of-concepts and this is likely to form a significant proportion of new projects for the data science team. But we are implemented in implementing industry best-practices along all elements of the product life cycle.
18. Can you provide a list of the internal and external stakeholders and partner organisations that either provide or consume the DIT data / insights?
The-Data-Science-team-will-sit-within-the-Digital, Data-and-Technology (DDaT) function. This-is-an-enabling-function-for-the-entire-department. DDaT-is-organised-into-various-portfolios-responsible-for-different-areas-of -work-such-as-export and investment and the wider trading environment.
Stakeholders will often be internal, such as profession leads whose work may be enhanced by data science or other product teams who spot an opportunity for data science applications in their product. It may also include policy colleagues who rely on particular reports or data to help them shape policy.
Key external stakeholders include HMRC, who rely on tariff data being delivered correctly into customs systems.
The department also works closely with other departments on various policy areas and ad-hoc data science support is sometimes provided.
19. Will the data insights feed to any automatic decision triggers or primarily to supplement information to aid policy direction or changes?
Any insights extracted from data will likely feed into a mixture of scenarios, including aiding and informing decision makers, enhancing DIT's products and services and providing basis for certain areas of policy direction. The core focus will be around enhancing DIT's digital products and services. For example, this could include improving how DIT deal with enquiries, DITs website content or how the department identifies and prioritises potential exporters and investors.
20. Does the authority engage adjacent teams to provide use case or user research. Alternatively will this responsibility be required of the new data science team?
Within the department, there are dedicated user researchers that help frame and define user needs. For most new projects, a multidisciplinary team including user researchers would be employed. For this innovation lab, some informal stakeholder engagement will be required from time to time in order to develop requirements but in general any user research will not be the responsibility of the data science team. Although Data Scientists with a strong appreciation of user needs will certainly be of clear benefit to DIT.
21. Is the full budget of the project going towards this bid or will there be additional phases?
There is no further Data Science procurement planned by DDaT in DIT. However there may be one if needed in the future. There is no guarantee that the full amount will be spent under this contract. Suitable projects for the Innovation lab need to be agreed on an individual basis between DIT and the supplier.
22. What is the process for awarding the SOW?
Each SoW is agreed on an individual basis for a fixed period of time. DIT and the supplier will agree a project that is suitable for a package of work. Timescales and deliverables are agreed up front.
23. What is the expected size and skillset of the team?
It is expected than the focus will be on highly skilled data science specialists who can set the direction in terms of application of data science. As the Data Science team has now been built up we do not anticipate significant need for basic or intermediate level data science resource. Resource need will vary by project and could range from Senior Data Scientist to lead within a product team on a particular feature or oversee other data scientists, to a full team of 4-6 to work on a time limited project.
24. Is there an incumbent supplier?
No, the previous contract has concluded. The previous supplier was PA Consulting Services Ltd.
25. Are there any other third parties involved that could potentially cause conflicts of interest?
No.
26. We note that the team operates under hybrid working arrangements could you please elaborate on the expectation for time to be spent within DIT offices?
For staff supplied via this contract there is no absolute requirement to come to the office. The civil Servants you will be working with have to be in the office 2-3 days a week. However in order to deliver effectively on this contract some attendance in the office will be beneficial, especially for key workshops and to build links across the DDaT function.
27. If successful following down selection to the top 3 suppliers, what are the timescales for the release of the full invitation to tender and the deadline for issuing full proposals?
This is an indicative timetable, that is subject to change.
Request for written proposals: Thursday 10th November
Deadline for proposal submissions: Thursday 24th November
Presentations: 6th-7th December
28. Can you please confirm if there is an incumbent who is delivering Data Science work within your innovation lab at present?
No, the previous contract has concluded. The previous supplier was PA Consulting Services Ltd.
29. Can you please confirm the approximate breakdown you envisage between supplying full teams vs individuals?
This is dependent on the projects that emerge, but also on the types of project the successful supplier can come up with. We are looking to a certain extent for the supplier to take the lead in terms of direction of innovation and data science.
30. Can you please confirm if DIT will sponsor SC clearance?
We are unable to sponsor SC clearance.
31. How far in advance will the statements of work be agreed? What will be the team stand up period? Do you propose to staff with full or part time roles?
This will depend on the project. Typically projects have been agreed for 2-3 months at a time, however some projects have been longer or shorter than this. Typically projects are agreed as need with work commencing immediately or in the following weeks.
32. It mentioned that some projects need further development – do you have a view on what type of projects they are?
Some of the key projects that DDaT data science has underway that the innovation lab could input into include:
-Company matching
-Data Science to support content delivery
-Lead generation
-Automated checking of Tariff and FTA data.
33. Are you able to outline the governance process for data science projects within DIT?
All DDaT products and services go through service assessment against the CDDO service standard, and teams are expected to work according to the standards set out here. Access to each dataset is controlled as the dataset level by dataset owners at the dataset level. For new tools and novel applications of data, projects will also have to go through our Information Risk assurance board and our data ethics board.
34. Are you able to give more details on the external users for these services?
The bulk of Data Science work has been internally facing. However Data Science have to date provided input on the following externally facing DIT services:
- Content at https://www.gov.uk/government/organisations/department-for-international-trade
- DIT's Tariff info at: https://www.gov.uk/trade-tariff
- https://www.great.gov.uk/
35. Please can you provide an organisation chart for the DIT and in particular for the Digital Data and Technology Directorate and specifically the existing Data Science team, indicating the number of people in the team and their roles?
This is not relevant to this requirement, DIT have provided a summary of the data science team above.
36. A Presentation is mentioned, as an ‘added assessment method’; please can you provide information as to when this is likely to be, and whether there is a specific scope, duration and format to this?
This is an indicative timetable, that is subject to change.
Request for written proposals: Thursday 10th November
Deadline for proposal submissions: Thursday 24th November
Presentations: 6th-7th December.
37. Are there any specific developments that have already identified as delivery priorities for the early months of the Data Science Innovation Lab, which would therefore be appropriate to share at high-level with all bidders ?
Our key workstreams at the moment:
-Company matching
-Data Science to support content delivery and enquiry management
-Lead generation
-Automated checking of Tariff and FTA data

These are subject to change depending on wider product priorities in DDaT as a whole.
38. Please can you confirm a what stage of the application suppliers will be asked for a written proposal?
This is an indicative timetable, that is subject to change.
Request for written proposals: Thursday 10th November
Deadline for proposal submissions: Thursday 24th November
Presentations: 6th-7th December.
39. Is there a need to transfer delivery responsibility from the incumbent or existing DDaT DIT team members as part of the work, or will each engagement be for new requests?
No. Work with the previous supplier has finished. The Innovation lab will work with the existing civil servant team in order to deliver data science outputs. The purpose of the innovation lab is to provide additional capacity and capability around for specific innovative applications of data science.
40. Is there any forecasted transfer of key resources (incumbent, contractor) to the successful supplier?
No, all of the previous work has been documented and ownership has been transferred to the core civil servant data science team.
41. Can you provide further details on what data visualisation, data science technologies and languages are in use?
Within reason there is scope to use any standard open source libraries. For Data Visualisations our data platform (Data Workspace) has QuickSIght and Superset, as well as enabling more custom visualisations using tools like d3 and streamlit.
Most of the key ML libraries have been used (e.g, XGBoost, spaCy...) in the course of the work of the data science team and there is scope to use other libraries where necessary.
42. Can you confirm the size, capacity and roles of the DiT DDaT Data Science team?
Head of Data Science. 5X Senior Data Scientist 3 X Data Scientist 2 X Junior Data Scientist.
We also have a separate teams for Data Analysis, Data Engineering and Data Architecture as well as other DDaT professions. All professions work closely together in multi-disciplinary teams.
43. Will the department offer the ability to sponsor further clearance for additional team members during the course of the engagement?
No unfortunately we cannot sponsor the SC clearance process.
44. Will there be a need for data engineering/ architecture team members or are data in a well structured environment?
Data engineering and architecture resource is provided in a different team, and is not a core part of what we require in this contract. However, Data Scientists with strong experience of Data Engineering concepts and technologies will be able to make a strong contribution and will aid implementation and application of data science in the team.
45. Dear Team, in the section 'How suppliers will be evaluated' the statement says 'All suppliers will be asked to provide a written proposal' and then 'how many suppliers to evaluate 3' - please can you confirm how many will be invited to provide a proposal?
3 suppliers will be asked to provide a written proposal in Round 2.
46. How many FTE’s would you anticipate needing on average?
We anticipate needing 2-4 data scientists at a time on this project. On occasion we may have a need for more resource. However what we are primarily looking for on this project is strong thought leadership and direction in order to drive innovation in data science and help build the data science capability in DDaT.
47. For a typical SoW what length of time would you expect a team to be engaged for and how large a team would you need? If augmenting an existing team, how many are needed from the supplier under this contract?
This obviously varies but we would expect each SoW to typically be for 2-3 months are require between 2-6 people. The lower end would be to supplement an existing team, while the higher end of this team would be to deal with a new emerging problem as a standalone team.
48. In your existing data science team, what is the platform of choice (not the cloud provider – AWS, the actual data science framework) and what are the tools you typically use? Is there a desire to migrate towards particular tools and techniques?
We are open to using any open source tools. The bulk of our work is caried out in python notebooks managed on our departmental cloud data platform. In terms of ML libraries we have used all the standard ones such as TensorFlow, spaCy, and XGboost among others. We are open to exploring other technologies subject to them being able to work within our technical architecture. To productionise models we use other tools such as Docker, Airflow and MLflow.
49. Have you already held any form of ‘meet the buyer’ webinar for questions and answers from suppliers? If there is a recording can it be made available? If you haven’t held one, would you consider holding one?
No we have not held any market engagement webinars, and are not planning to do so.