• String
  • Posts
  • Events aggregator, EduTech ecosystem builder and AI remarks/ referral co-pilot

Events aggregator, EduTech ecosystem builder and AI remarks/ referral co-pilot

String Q3 2022 to Present

String has gone through a few iterations.

Q3-Q4 2022 - String Explorations, pivoting from students to focus on teacher development

There were 2 main products:

  1. Events aggregator - premise: students and teachers learn by organizing and partaking in events that focused on tech skills.

  2. Two-sided marketplace for EduTech problem providers and builders - at this point, we actively spoke to teachers and gathered problem statements. Premise: there are genuine constraints to teachers solving their own problems which obliged creating an environment.

If I can’t, who should I turn to? What environment might better support educators?


  • This was the last batch of student ‘interns’ we took from SAJC for their annual work shadowing program.

  • String heavily pivoted from a student-centric model to focus on teacher exposure to emerging technology

  • Onboarded/ engaged National Institute of Education (NIE) Office of Teacher Education (OTE) to explore ways to provide such exposure upstream to (revisiting in Sep 2023)

  • Tech: Primarily Notion as MVP, Figma and some attempt to build and launch in La Ravel but failed due to lack of dedicated engineering effort

Q1 2023 - String as a tech ecosystem builder, achieved via ‘simplify’, ‘scale’, ‘structure’. String was pitched as a hackathon idea under Open Government Products’ annual internal hackathon, Hack for Public Good.

  1. Simplify - continued to run EduTech exposure events under the banner Tech Talk for Teachers.

  2. Scale - we take good, scalable scripts out of local computers to encourage adoption across schools via a GitHub organisation, String, modeled after other open source platforms like NaasAI. But not all teachers are familiar with GitHub, so we also have a more beginner-friendly aggregator. We also aggressively experiment and put experimental tech like CherGPT into the hands of teachers.

  3. Structure - we tried pursuing partnerships with Friends of Figma (Educators), Google, among other collaborations.

Q2 2023 - String as a tech ecosystem builder (continued)

  1. Simplify - we concluded our 30th session of Tech Talk for Teachers in June 2023, impacting a cumulative total of 400 teachers.

  2. Scale - CherGPT had a mini-existential crisis but it took a different turn as the base code of prototypes by MOE HQ arms like AI Learning Companion (AILC), Metacognition Team and made its way to couple more schools. Cumulatively we had >2000 unique users, >40,000 engagements with AI. CherGPT had a surprise mention by Education Minister Chan Chun Sing in his keynot address at MOE’s ExCEL Festival 2023.

    • We continued more generative AI experiments with a rough MVP of Remarks Co-Pilot, deployed using a Google Sheets connector

    • Concurrently efforts for Referral Co-Pilot with Hwa Chong continued

      We know whatever tools/ product we put out are not perfect, but we learnt so much just trying things together

  3. Structure - we hosted our second in-person meetup at Google APAC HQ, attended by a total of over 50 educators as well as staff from MOE HQ, unveiled more experimental tech products with a focus on the qualitative writing space while negotiating for developer perks.

Our meetup on 6 June 2023 - did we mention that you also can have Google Cloud credits when you develop with String?

Q3 2023 - String as a tech ecosystem builder, focusing on scaling our own tech by actively going to schools to co-create prototypes

Remarks Co-pilot, a generative AI writing assistant tool for teachers to write short qualitative remarks, has been tested in a few environments:

  1. Microsoft Teams, tested with Crescent Girls School via Norman, Eddie, Yew Li, Wen Yeow and Team

  2. Google Sheets, actively departed by Wesley in ACS; first iterated by Mathew previously from Hai Sing, currently at NJC

  3. Python/ Streamlit, tested by Lance from Serangoon Secondary, Adrian from CHIJ Toa Payoh, among other schools

While ambitious, we might just launch our own full-stack version (soon).

The learning has been rich in co-creating tools and developing a much better understanding of how remarks/ testimonial writing teams function and what commonalities in workflows exist.

We can never meaningfully scale if we build things alone. String attempts to knowledge share via platforms like these and actively working with schools’ ICT leaders and any teacher curious about adopting and/or building tech

In the coming articles, we will be diving deeper into unpacking the problem of qualitative writing and how the String team, working alongside teachers, have created tooling to reduce admin time to have that first draft to work off.

Every conversation I have with teachers reminds me that there is ever so much to do.

But here is to returning that few hours in such repetitive tasks without displacing authentic representations of our students - that is our goal for term 3/Q3 for educators. Watch this space :)


  1. How does ‘scaling’ work for educators? Does that mean they learn how to build themselves or they work with the tech peeps in quick prototyping?

    Teachers build EduTech typically via basic tooling (not full-stack apps per se, usually no backend). Simple tooling is done on Google App Script and vanilla javascript. Another increasingly popular mode of deploying code is Python via Streamlit.

    Ideally, they work with technical people. String concurrently pursues a community of EduTech enthusiasts. There are a number of more technically inclined teachers sharing their experience with others. But there are clearly gaps. One gap that String hasn’t been able to patch is collaborative code and version control. More members now host and deploy code via Github. But concepts like branching strategy, containerization, code revisions, having staging v UAT v production environments, testing, etc are not apparent. This is in large part because teacher tech is typically done in silos of deploying in local computers instead of actively collaborating and doing versioning across teams.

  2. What are the outcomes of ‘structure’ arm?

    String will organize 2 learning journeys this year for the teaching fraternity. The first was held at Google. The second, scheduled for Nov W3/4, will be at Meta.

    Some of our influence will also go upstream to the National Institute of Education (NIE) Office of Teacher Education (OTE).

    Ideally, we clutter bargain for developer perks for teacher-developers via such partnerships so the friction to deploy experimental tech is lower.

How can I participate in String?

String is an ecosystem builder - lots of opportunities to join regardless of your skill/ experience level. We have a couple of opportunities lined up:

  1. User Test our AI writing Suite - this signup link still point to the old GSheets prototype but we will keep you in the loop once we have the new product ready

  2. 2 Sep 2023: User Test for ExploreSG (run by a team at Open Government Products) - join us for an in-person field trip at Chinatown, keep a lookout on SG LDC and Discord 

  3. Share this blog/ newsletter with like-minded educators - we have 2 more articles on workflow automation and integration in the context of education coming up in the next 2 weeks, covering both Assessment for Learning as well as reduction of long-form text writing effort on the part of teachers

  4. Bring String to your school - we are most happy to chat with your ICT/ Professional Development units to make learning for teachers fun, relatable, and applicable.

  5. Sep 2023: Join us at the upcoming SG LDC Virtual Meet (:

  6. Nov 2023: Join us for our last in-person String meetup/ learning Journey at Meta, details to come also Discord 

Info overload? 1 link with everything String: go.gov.sg/stringme