Women of Silicon Roundabout: Day 2
See my last blog for my write-up of Day 1See my last blog for my write-up of Day 1
Now, onto day 2!
Keynote 1 - A Journey to an Inclusive Net-Zero Future
The first keynote was given by a panel of women from OVO Energy - Shraddha Nayak, Sara Gotham, Bronwyn Barnett and Nicola Spanton. They had some really interesting conversations around:
- The differences between mentors, coaches and sponsors.
- Changing messaging for women moving into leadership roles. Instead of encouraging them to adopt "male" traits, celebrating different leadership styles.
- The need to raise awareness about the breadth and variety of opportunities within tech.
- Positive bias.
- And, supporting neurodiversity in the workplace.
On this last point, they discussed:
- The ways it can be an asset.
- The importance of specific coaching for managers who are managing people who are neurodivergent.
- Coaching for people who are neurodivergent around helping them work out what they need to succeed.
- And, how AI can be used to help people with dyslexia.
Keynote 2 - How Tech Is Flipping the Energy Game
The second keynote was given by a Nina Moger (Transformation Director, Kraken), and it was excellent. She was incredibly passionate and inspiring, and had some great insights into how the energy sector needs to change to combat climate change.
The energy industry is historically slow to change, and Nina highlighted how we no longer have time for that. She talked about the different ways that she has tried to affect this change in her life up to this point (through government, through academia, and now in the private sector). And how she now feels that she is in the right place to start to make a difference - even at this late hour in the fight against soaring global temperatures. It was both a message of doom (with some horrifying graphs) but also one of excitement and hope.
She talked about three ways the energy sector is changing:
- The national grid is moving away from fossil fuels
- The shape of consumer demand is changing (rooftop solar panels, electric heating, charging of cars overnight)
- There is suddenly a huge wealth of data, from smart meters and other devices, and it's predicted that there will be another 200 million devices by 2030. This is a huge, live, dataset...
A Reaqtor shaped problem?
This last point particularly piqued my interest because it reminded me a lot of a project that endjin have been working on for almost 10 years - Reaqtor.
Reaqtor is a reactive query engine built for reliably handling large volumes of real-time data. It is built on Rx.NET, and is the technology behind Microsoft's Cortana. When endjin first got involved in the project, it was relating to a conversation with TalkTalk around how they could support the maintenance and monitoring of millions of devices at once. The type of problem that these energy companies are now dealing with - with large volumes of live data and lots of actors / streams, are exactly the sort of thing Reaqtor is perfect for. Not only that, but it's extremely low compute - it can even be run on a raspberry pi - meaning that you could push a lot of this real-time computation closer to the data sources...
The talk then went on to cover how changing consumer demand presents an opportunity - to affect it.
In order to reduce emissions, energy companies need to match times of high demand with times where there is high green production - e.g. it's sunny, windy, etc. across the country. Nina talked a lot about the importance of flipping the mindset of production needing to match demand, and instead about how we can help demand match production. This is important in terms of renewable energy, but also more generally if we are going to stop depleting the planet's resources!
Kraken is an offshoot of Octopus Energy and this is essentially their goal. They have started to achieve this with Octopus, but also hoping to branch out to other energy companies. They are doing this in a few ways, by incentivising people to use:
- More power when there is local power being generated (by a local turbine or solar panels etc) - by giving them a discount
- Less power when there is more stress on the system - by giving them (financial) rewards
And, they also have some plans around providing people with free energy if they are able to manage their entire home for them (it begins to sound like something out of sci-fi, but I can definitely see the benefit of having more control over usage at a larger scale!).
This was also interesting in terms of what we've been thinking about recently here at endjin. James Broome recently posted a blog series around reducing environmental impact when running workloads in Azure. Following on from this, we had a discussion around how if these things are going to be done on a large scale (rather than everyone developing strategies independently), then it needs to be managed centrally. In terms of pipelines, you can imagine that if everyone implemented a solution to move their compute as the sun came out, then you could suddenly get huge and unpredictable load on certain datacentres. However, if Microsoft managed this centrally they could provide an "eco" tier for those who don't mind when or where their jobs are run, and organise the compute to have the least environmental impact. This is essentially what Kraken are trying to do, but instead in the energy sector itself.
She then answered some questions around how she manages to stay positive and talked about how it's important to be able to balance hope and outrage at the same time. I often struggle with this in finding the right balance - outrage is important fuel for action, but can also easily transition into a sense of disillusionment and powerlessness. The importance of hope cannot be understated.
Finally, she answered some questions about AI and the power usage, and how it was a matter of ensuring that the benefits far outstripped any energy that was used - and about how clearly all datacentres need to be run on renewables.
Overall, the talk was brilliant and I left feeling very inspired!
Enabling a more sustainable Internet of Everything
This was a workshop session, run by Catherine Ramsdale (Pragmatic Semiconductors). Pragmatic have invented a new type of chip using thin-film semiconductors rather than silicon. They are lightweight and flexible, and made at a fraction of the cost and environmental impact of silicon chips (due to less steps in manufacturing, less harmful chemicals, less water used, etc.).
The chips themselves are passive and hold a small amount of data, but can be scanned by a receiver (which could conceivably be any phone!). The technology is still relatively new, but they are starting to think about different applications as it progresses. At the moment you have to scan the chip at close range to retrieve the data / know where it is (much like scanning in/out of the London underground system). But they are working on larger range fields for easier detection.
They are currently thinking about using it in waste management. They want to track products as they move through manufacturing and then through recycling centres, and other waste handling processes. This would enable us to understand, in more detail, how waste was being managed and to adapt these processes is the best way possible.
They are also thinking about reducing food wastage by attaching a sensor that senses whether food has gone off, and therefore meaning that food doesn't need to be thrown away before it is actually inedible. Currently use-by dates necessarily add in a large amount of room for error - so are very conservative in their estimates.
The workshop was mainly a big brainstorming session for use cases. And there were some ideas floated around tracking sustainability of supply chains - around fabrics (which would make use of the chip flexibility). People also came up with use cases around tracking when medication refills are necessary, and various other medical applications.
I had a couple of ideas over the course of the session:
Tracking animal migration patterns
One of the big challenges in understanding ecosystems around the world is understanding how animals move around. For example: I was shocked to discover last year that, despite them being one of the largest animals in our ocean, we have little idea of the migration patterns of whale sharks. This is also true for birds, and many land animals. A challenge with tracking these things is often the large cost of tracking devices, and the need to have devices that won't impede the animals in any way. I thought this could be a great application for these low cost and lightweight chips (once the field of detection increases!).
Tracking carbon filter usage
Another thing that I was thinking about was the possibility of using them to detect how much water has flowed through carbon filters, when they are used for generating clean drinking water. Carbon filters filter out bacteria, parasites, viruses, and harmful chemicals. But, they present a challenge in that, as they are used, the carbon degrades. This means that they can only be used for a given amount of water - for example my lifestraw bottle can filter 100L of water before the filter needs to be changed.
This is particularly challenging because there is no way of telling when a filter is at its end of life. I, for example, had a note on my phone where I attempted to keep track... But you can imagine this isn't the most reliable of solutions! And, this problem increases when you, for example, attach these filters to a tap as you lose the ability to track water usage at all.
Using these chips we could detect whether the carbon filter is still able to filter things correctly. This means you could increase the amount of time that filters could be used for, as they currently have to be conservative (just like with use by dates). You could also alert people when the filters need changing, and track where around the world filters need to be replaced (if you were, for example, a charity that works to provide clean drinking water). The fact that it's low cost is important because efforts to supply drinking water around the world are mainly run by NGOs, which obviously have huge budget constraints.
Finally, we talked about how this could enable further expansion of the internet of things, as it allows the connection of items that would otherwise be difficult - those that are lightweight, low cost, or flexible. It is also interesting to again link this to Reaqtor which, with its incredibly low-compute processing capabilities, could allow more processing in-situ of the data supplied in these situations.
Beyond Burnout: Thriving as a woman in tech
This session was given by Diana Foltean (Head of Data Sharing & Partnerships). She shared a lot about her own journey with burnout. She also discussed how her neurodivergence contributed to her reaching the point of burnout, and continues to affect her life at work.
She talked about the three stages of burnout:
- A huge amount of excitement and enthusiasm meaning you want to give 110%
- This giving way to stress
- And, finally, exhaustion
It was interesting to think about how burnout often comes from a place of caring; You want to do your best because things matter to you, but there's clearly a point at which that becomes unhealthy. She talked about how women, especially neurodivergent women, are more prone to this. I've had a lot of conversations in the past around the additional pressure that women in STEM feel to overcompensate and prove that they can keep up. It is not surprising that feeling this constant need to go the extra mile can easily lead to you reaching that place of exhaustion. But, this is even more true for people who are neurodivergent - who are trying to keep up in a world that is, even more so, not built for them.
She spoke about the neurodivergent spoons analogy - which is worth looking into if you're not familiar, but essentially it's the theory that people only have a certain amount of "spoons" that they can give in a day. These include "eating", "washing up", "meetings", "exercise", etc. And if people run out of spoons then they start to borrow tomorrow's spoons, and the next day's, and this build up of frantic trying to keep up is what leads to burnout.
And that, especially for neurodivergent people (but I think is worth keeping in mind for everyone!), it's important to understand which tasks are drainers, and which give you energy. This is person-specific - I may, for example, find extended periods of socialising exhausting, where others might feel envigourated by it! Understanding this means that you can manage yourself, and your work, much better.
She then talked about how companies have a responsibility under the equality act to provide support for people who are neurodivergent. This was an important point because I think that many people don't realise that they not only can, but have the legal right, to ask for this support.
This can look like:
- Flexible working - meaning that people can organise themselves in a way that works, not working for extended periods in one go, working later in the evening if that's better for them, etc.
- Providing coaching for managers of neurodiverse people
- Providing coaching / therapy to help people understand their drainers and energizers
- Extra 1:1 support
- Administrative support
- And much more!
Finally, she talked about how she necessarily took a sabbatical from work in order to recover from burnout. She spent the first two months doing literally nothing to recover. Then, the next two starting to heal. Then the final two did some travelling to recapture some joy. And I thought that this was a really powerful message, both around the fact that it can get on top of you to the point you have to literally do nothing for two months, but also the fact that you have to recapture that joy in order to be ready to return!
Everyone can write python, but not everyone can write good python
The final session I went to was around python engineering practices. It was given by a Meghan Avery (Senior ML Engineer, AmEx) and it was interesting to hear that a lot of their backend is python-based. Similarly she said that YouTube and Instagram have entirely python-based back ends - who knew?!
She talked about the common issues with python - which I think that many people who have worked in the python ecosystem are aware of, around:
- Package and version management
- Dynamic typing (coming from a C# background, python feels to me a bit like the Wild West!)
- Lack of compile time errors
- Different programming styles
- Lack of testing and quality checks
And then talked about mitigation of some of these issues:
- Type hinting - which you can use to, not enforce, but give information around types in your code
- Using linters to enforce coding standards, such as PEP8 Google standards (we are using Flake8 here at endjin!).
- Making decisions on programming style as a team. For example, whether to use procedural or OO programming, and then using one or the other as standard across your codebase.
- Enforcing code coverage using PyTest - I have mixed feelings around this as enforcing code coverage doesn't always actually improve the testing of code... However, it can be useful in making people actually think about testing and what has and hasn't been validated.
And that's a wrap!
Overall, I really enjoyed the second day of the conference. I saw some excellent talks - both technical and non-technical, and I really enjoyed getting fully involved in the brainstorming during the workshop. There were some interesting links into things we've been working on here at endjin - especially around Reaqtor - and we've had some great conversations as a result.
Again, thank you to everyone who spoke, and the event organisers for bringing us all together!