If you work in computer vision or machine learning, here is a scholarship to pause on. Monash University partnered with Aervision Pty Ltd, an Australian AI company, to fund a (fully paid) Industry PhD Scholarship that relates to privacy-preserving AI for human-centric video analytics. You will work at Monash’s Clayton campus in the Department of Data Science & AI in the Faculty of Information Technology, and the whole thing is supported by the Australian Government National Industry PhD Program.
What makes this one different from the standard PhD listing? It’s different from just doing theory in a lab. You will divide your time between Monash’s research environment and Aervision’s product team using real industry data and viewing your research actually shipped into a commercial platform. Not everyone who is doing a doctoral degree gets that experience.
What This Opportunity Actually Involves
Cameras are everywhere now and video analytics software is doing more of the heavy lifting in reading what’s happening in front of them. However, the problem is that reading human behaviour accurately usually means keeping identifiable footage around, and that clashes hard with privacy expectations and with the law. The systems must also run in real time on whatever hardware is sitting on-site.
This PhD project tackles that head-on. The goal is to build foundation models that stay accurate at reading human behaviour, adapt to new locations without needing a full retrain, and protect people’s identities by design rather than as an afterthought.
You’ll work under Dr Deval Mehta, splitting your time between Monash’s academic research setting and Aervision’s R&D team. There you get to peruse a setup that gives you a front-row seat to how research ideas turn into shipped software.
| Detail | Information |
|---|---|
| Job Reference | 696616 |
| Campus | Clayton campus, Monash University |
| Employment Type | Full-time |
| Duration | Up to 3.5 years full-time PhD study |
| Supervisor | Dr Deval Mehta |
| Department | Data Science & AI, Faculty of Information Technology |
| Industry Partner | Aervision Pty Ltd |
| Funder | Australian Government National Industry PhD Program |
| Deadline | Sunday, 30 August 2026, 11:55pm AEST |
How Much Does It Pay? Funding Breakdown
The funding you receive as a PhD student isn’t a single flat amount paid from a single source. Here’s what you get:
- A Research Living Allowance of $37,145 AUD a year (2026 rate, indexed annually) for your whole PhD.
- A Faculty of Information Technology Tuition Fee Scholarship for eligible international students.
- An Industry Top-up Scholarship of $10,000 a year, paid for by Aervision.
- FIT Candidature Funding of $4,000, covering the full length of your candidature.
- A one-off travel grant of up to $1,265 from the Monash Graduate Research Office.
- A Top-up Government Scholarship worth $7,135 a year.
Add it all up and you’re looking at a funding package that goes well past a typical PhD stipend.
The Actual Issue Being Addressed by This PhD Study
Modern video analysis tools must accomplish three challenging tasks simultaneously: accurately interpret nuanced human behavior; keep people on camera anonymous; and function in real-time, either on deployed hardware or in the digital cloud.
The Real Problem This PhD Is Trying to Solve
Video analytics software today is being asked to do three hard things at once: read subtle human behaviour correctly, keep the people on camera anonymous, and run in real time on deployed hardware. Most systems trade one off against another. Push for accuracy, and you usually end up leaning on identifiable footage, which opens the door to privacy and compliance headaches.
Push for privacy instead, and accuracy tends to drop, or the model falls apart the moment it’s pointed at a new environment it wasn’t trained on.
This project aims to break that trade-off. The plan is to build AI models that generalise well and still read human behaviour reliably, but that work from anonymised representations of people instead of raw, identifiable video.
Research Areas You’ll Get to Explore
There’s a lot of room to move here. Expect your work to touch on:
- Human action and activity recognition
- Understanding behaviour at both the individual and group level
- Predicting behaviour over longer timeframes
- Multi-task learning and foundation model architecture
- Zero-shot and open-set recognition
- Privacy-preserving AI built on anonymised human data
- Getting AI models to run efficiently on edge devices
- Applied AI for security, healthcare, and smart-environment settings
You’ll be working with anonymised datasets supplied by Aervision, and everything you build gets tested directly on Aervision’s live commercial platforms. That’s a genuinely rare pipeline for a doctoral project as most PhDs never get validated against a real product before graduation. To top it, it also feeds into Australia’s wider effort to build AI that’s trustworthy and actually gets used commercially.
Who Fits This Role
Monash and Aervision want someone with a strong base in computer science, data science, or AI, plus real curiosity about both academic research and working alongside an industry team.
Skills You’ll Need
- A solid handle on deep learning and machine learning basics
- Computer vision experience
- Comfortable coding in Python
- Hands-on work with PyTorch or TensorFlow
- Strong maths and analytical thinking
Skills That Give You an Edge
None of these are dealbreakers if you’re missing them, but they’ll help your application stand out:
- Video understanding or human action recognition work
- Experience with vision-language or other foundation model setups
- Any prior privacy-preserving AI work
- Having trained large-scale models before
- Papers published at top venues like CVPR, ICCV, AAAI, or NeurIPS
Being a clear communicator matters too as you’ll need to explain your work to both academics and a commercial team.
Do You Meet the Academic Requirements?
You need to satisfy just one of these three routes into the PhD:
- A four-year (or longer) bachelor’s degree in a relevant field, including a research thesis or project, with grades equivalent to First Class Honours; or
- A master’s degree in a relevant field with a research component worth at least 25% of a full-time year, again at a First Class Honours equivalent; or
- Some mix of qualifications and professional experience that Monash’s Graduate Research Committee judges to be equivalent.
Monash also points to its commitment to diversity, fairness, and equity, and backs the gender equity goals laid out in the Athena SWAN Charter.
What You Get Beyond the Money
The stipend is only part of the deal. Here’s what else comes with the role:
- A government-funded Industry PhD Scholarship, which carries real prestige
- Ongoing, direct work alongside an Australian AI company doing real product development
- Access to large-scale industry datasets most academics never get near
- An industry placement that runs the entire length of your candidature
- Full use of Monash’s high-performance computing setup
- A genuine shot at publishing in CVPR, ICCV, AAAI, NeurIPS, and similar venues
- Formal training in how research gets commercialised and turned into a product
- Solid career prospects afterward, whether you head into academia or industry
If you’re weighing 3.5 years of your life against other PhD options, the mix of hands-on industry work plus a real publication pathway is what makes this one different from a standard university-only doctorate.
How to Apply: Two Stages
This role doesn’t run through the standard one-shot PhD application. It’s a two-stage process, built to check your research thinking and project fit before you put together a full application.
Stage 1: Send an Expression of Interest
Email Dr Deval Mehta directly at deval.mehta@monash.edu. Your EOI needs:
- An up-to-date CV
- Your academic transcripts
- A cover letter
- A draft research proposal, no more than 5 pages, laying out how you’d approach building privacy-preserving AI for human-centric video analytics
That research proposal carries the most weight in Stage 1. Use it to spell out what’s currently missing or broken in video analysis research, sketch a framework that addresses those gaps, and show where your approach brings something genuinely new. Also explain why you want to do a PhD in the Department of Data Science & AI, and why this particular project grabbed your attention.
Stage 2: Talk It Through, Then Apply Formally
If you clear Stage 1, you’ll get invited to talk through your research ideas one-on-one with Dr Deval Mehta. From there, if things go well, you’ll be guided through putting together and submitting your full PhD application.
The deadline sits at Sunday, 30 August 2026, 11:55pm AEST. However, here’s the catch: Monash isn’t waiting until then to start reviewing. Shortlisting and interviews kick off as soon as strong applications land, so getting your EOI in early gives you a real edge over waiting until the last week.
Who to Contact
- Enquiries: Dr Deval Mehta — deval.mehta@monash.edu
- Job Reference: 696616
- Campus: Clayton, Monash University
- Industry Partner: Aervision Pty Ltd
Bottom Line
It’s uncommon to find this combination of a fully government-funded industry PhD with oversight from an actual AI company. Should your research interests align with computer vision or foundation models, then this project is worth considering. Since Monash reviews applications on a rolling basis, don’t sit on this. Rather, get your research proposal moving now rather than waiting for the deadline to creep up.
Frequently Asked Questions
How much is this scholarship actually worth?
There’s no single number because it’s built from several funding streams. You get a $37,145 AUD Research Living Allowance per year (2026 rate, indexed annually), a Faculty of IT Tuition Fee Scholarship if you’re an eligible international student, a $10,000-a-year Industry Top-up from Aervision, $4,000 in FIT Candidature Funding across your whole PhD, a one-off travel grant up to $1,265, and a $7,135-a-year Top-up Government Scholarship. Stack those together and it’s a well-funded, multi-source package.
Who exactly is Aervision, and what’s their role in this?
Aervision Pty Ltd is an Australian AI company building video analytics tools for critical infrastructure, public safety, and smart environments. As the PhD candidate, you’ll spend part of your candidature working directly inside their research and development team, getting real exposure to how commercial AI products get built and shipped.
How long is the PhD funded for?
Up to 3.5 years of full-time study for a Research Doctorate at Monash.
What do I need academically to even apply?
One of three things: a four-year bachelor’s degree with a research thesis and First Class Honours-equivalent results, a master’s with a substantial research component and the same honours-equivalent average, or a combination of qualifications and professional experience that Monash’s Graduate Research Committee accepts as equivalent.
What skills do I actually need going in?
You’ll need solid deep learning and machine learning fundamentals, computer vision experience, Python skills, and comfort with PyTorch or TensorFlow, plus strong analytical and maths ability. It’s a bonus, not a requirement, if you’ve also worked on video understanding, foundation models, privacy-preserving AI, large-scale training, or published at CVPR, ICCV, AAAI, or NeurIPS.
What’s the actual process for applying?
Two stages. First, email an Expression of Interest to Dr Deval Mehta at deval.mehta@monash.edu with your CV, transcripts, a cover letter, and a research proposal capped at five pages. If that lands well, you’ll move to Stage 2 which is a direct conversation with Dr Mehta about your research ideas, followed by support putting together your full formal application.
When’s the cutoff to apply?
That’s Sunday, 30 August 2026, at 11:55pm AEST. That said, Monash starts shortlisting and interviewing the moment strong applications come in, so don’t wait around for the deadline. Consider applying as early as you can.
What will my research actually focus on day to day?
You’ll be building privacy-preserving foundation models for reading human behaviour on video. That covers action and activity recognition, individual and group behaviour analysis, long-term behaviour prediction, multi-task learning, zero-shot and open-set recognition, anonymised-representation privacy techniques, edge-device deployment, and applied systems for security, healthcare, and smart environments.
Who do I reach out to if I’ve still got questions?
Dr Deval Mehta is your point of contact for anything related to this project or the application process. You can email him at deval.mehta@monash.edu.