---
title: SHU Interview - Presentation 1 and 2
tags: interview
description: View the slide with "Slide Mode".
---

## SHU Smart Computing
### Interview Presentations

<!-- Put the link to this slide here so people can follow -->
Martin Callaghan
slides: https://bit.ly/mtc-shu-1

---

## Who am I?

- Martin Callaghan **Candidate 6**
- Senior Research Software Engineer
- I work with researchers to integrate software with (big) hardware & Cloud with (big) data with (big) AI
- Lots of teaching, training and staff development

---

# Presentation 1:

The challenges of **implementing** a smart **application** that requires the **integration** of **multiple** hardware, software and data components.

---

### Point to note

*Theres a world of difference between developing IoT/Smart solutions in a research or industrial context vs. teaching about it, developing skills and understanding context*

Note:
We want to make the learning and assessment experience authentic, but with a low barrier to entry to give our learners opportunities for success.

---

### Introduction

We can see this as an end-to-end Data Science and Software Engineering problem.

We’re attempting to:

**Combine data from the real world with AI (ML and DL) with human insight to drive Smarter Decisions**

Note:

---


---

### Types of Challenge

I'll categorise these as:

**S**oft
**H**ard

challenges

---


---

### Methodologies

* Robust (business) analysis techniques
* Agile development
* ‘Service Catalogue’ of Res/Dev/Cloud tools and techniques
* Leading to a:

### **Clear problem statement**

---


---

### Training and Development

When I work with a "customer", I don’t know what they know.

We need to ensure we understand each other's skillbase and support each other in developing new skills.

Note:
Authenticity again. Encourage our learners to see this as a real engagement, as a customer with a business need and not an abstract learning or research problem.

---

### My RSE role is to:

* Explore, Coach and Mentor set of skills and behaviours.
* So that we *coach-in* the same to our students.

---


---

### Social, Environmental, Ethical

**IoT/Smart Things** is a direct proxy for AI and has the potential to help us meet *Grand Challenges*
(in **Health**, **Climate** and **Enriching Culture**)

---

### Social, Environmental, Ethical

Also potential for **misuse** so [understand principles of fairness](https://www.microsoft.com/en-us/ai/responsible-ai?activetab=pivot1%3aprimaryr6)

---


---

### HW and Sensor platform

* Something that’s usable, motivational and heading towards ‘industry standard’.
* **Avoid the trivial.**

* Raspberry Pi (Pico, Nano...)
* Arduino & ESP8266-based platforms
* NVidia Jetson Nano (GPU for **Edge** compute)

---

### HW and Sensor platform

**Key requirements:**

* Robust & Easy to program
* Support for Analogue and Digital sensors
* Easy to abstract away complexity to make it teachable
* Connectivity (WiFi, BLE, 4G/5G)

---


---


---


---

### But hang on...

(Almost) everyone has a sensor platform with them all the time. How can we leverage those?

* Smartphones
* Smart Watches
* Connected Home Devices
* Connected fitness and health devices

---


---

### IoT Data Hub

Our smart devices have to send their data somewhere (the **middle bit** of our pipeline).

Look at:

* [Azure IoT Hub](https://azure.microsoft.com/en-gb/services/iot-hub/)
* [Particle.io](https://particle.io)
* [Temboo](https://temboo.com)

---

### IoT Data Hub

We're choosing between:

* Enterprise
* Industrial
* Hobbyist/Lo-code

---

### IoT Data Hub

BUT Enterprise Platforms also provide a wide range of other PaaS, IaaS and SaaS including **Cognitive Services** that makes architecting a solution much easier.

**Let's teach what's going to be useful**

---


---

### Downstream analytics

* What are we analysing and why?
* What decisions are we making?
* What are the consequences of these decisions?
* What tools and techniques (ML, DL) do we need to explore?
* How are those decisions communicated

---

### Example

Fitbit Data for Health

*Does bad weather activity levels for University students*

* Collect fitbit activity data for a cohort
* What data acts as a proxy for activity?
* How do we rank this data?
* What statistical tools and techniques do we need to employ?

---

### Questions?

---

### Presentation 2:

**Teaching** a second year undergraduate class about implementing a smart application that requires the **integration** of multiple hardware, software and data components.

---

### Design Philosophy

**Reverse Instructional Design**

* Learning Objectives
* Assessment and understanding
* Learning Activities
* Teaching Activities

---

### Teaching approach

* Learning activities to address each part of IoT pipeline
* Motivational examples
* Combine different stages
* Lots of formative assessment
* Final assessment
* via series of *Learning Sprints*

---

### EdTech

* Still in weird times so
* ... make use of
* Jupyter Notebooks
* BinderHub
* Cloud tools
* Github Classroom

---


Note:
Agile underpins most working practices in industry. Let's experiment with this approach
I'd see second year as the consolidation and practice layers of Bloom's taxonomy giving the breadth ready for depth in Yr3 and final year project.

---

### Weeks 1-5

|Week|Content|
|----|-------|
|Week 1| Introducing Things |
|Week 2| The role of AI |
|Week 3| [Investigating sensor data](https://colab.research.google.com/drive/1IK_VZtLJhTF6nyt3mFqUiBSNCdQ8RVVw?usp=sharing) |
|Week 4| Making decisions with data |
|Week 5| Sensors and devices for IoT solutions|

---

### Weeks 6-10

|Week|Content|
|----|-------|
|Week 6 | [IoT and Cognitive Services](https://drive.google.com/file/d/1sGfMHw6XFSHrJQPpj119mxIcbcJkvnx5/view?usp=sharing)|
|Week 7 | Connecting sensors to the Cloud|
|Week 8 | Real time AI in the Cloud|
|Week 9 | Final Assessment |
|Week 10| Final Assessment |

---

## Questions