Teaching Assistant
TARA
📍 Remote (Global) 🕔 0.25 FTE
💰$80 AUD/year ⏳ 23/01/2026
The Technical Alignment Research Accelerator (TARA) is a 14-week program launching in March 2026, designed to upskill technical APAC talent in AI safety. Using the ARENA curriculum, TARA covers advanced ML topics including transformer architectures, RL, mechanistic interpretability, sparse autoencoders, and model evaluation techniques.
As a Teaching Assistant, you'll work remotely to guide approximately 26 participants across two city cohorts within your assigned timezone cluster. All instruction and support is delivered online - you don't need to be physically present in any of the cities. You can be based anywhere in the world, as long as you're available during your cluster's Saturday session hours. You'll be part of a team of 3-4 TAs, each responsible for one of three clusters:
Cluster 1: Singapore + Taipei — Saturdays 10:00 AM - 5:30 PM SGT (UTC+8)
Cluster 2: Manila + Tokyo — Saturdays 10:00 AM PHT (UTC+8) / 11:00 AM JST (UTC+9) (~7.5 hours)
Cluster 3: Sydney + Melbourne + Brisbane — Saturdays 9:30 AM - 5:00 PM AEST (UTC+10/+11)
All cohorts run synchronised (everyone does Week 1 together, Week 5 together, projects together), so you can support each other with lecture prep, queue management, and project feedback.
The role involves leading Saturday sessions and providing flexible remote support. Total time commitment is approximately 10 hours weekly: 7.5 hours on Saturdays plus 2.5 hours of weekday support and preparation.
Course Structure & Weekly Schedule
Each Saturday begins with you introducing new technical concepts, followed by pair programming sessions. The standard Saturday schedule (~7.5 hours):
Hour 1: Lecture introducing the week's topics
Hours 2-3: Pair programming with remote/online TA support
Midday: Lunch
Hours 4-7: Continued pair programming with TA support
End of day: Daily download - participants share progress and predict challenges for the coming week
During the week, participants work independently on the material while you provide asynchronous Slack support (~2.5 hours).
Core Responsibilities
Saturday session delivery
Lead 1-hour lectures on weekly curriculum topics covering topics such as RL, transformer architectures, mechanistic interpretability, sparse autoencoders, and model evaluations
Proactively check in with pairs via Slack to see if they need help
Jump on Zoom calls or Slack huddles to work through problems as they arise
Encourage "learning in public" - when you help someone solve a problem, post the resolution so others can benefit
Weekday support
Offer asynchronous support via Slack for participants working independently
Help resolve environment setup issues, compute access problems, and technical blockers
Assessment and feedback
Review participant project proposals during the final curriculum phase (Weeks 8-11)
Provide technical feedback on projects (Weeks 12-14)
Team collaboration
Participate in weekly TA check-ins
Step in for other TAs if needed - our multi-TA model provides backup coverage
Share learnings and resources across clusters
Qualifications
Required
Completed most or all of the ARENA curriculum
Strong Python and PyTorch skills
Strong grasp of RL, transformer architectures, mechanistic interpretability, sparse autoencoders, and model evaluation
Experience explaining complex technical concepts
Ability to mentor in programming/ML
Patient and encouraging teaching style
Proactive communication habits
Genuine interest in AI safety
Available for the icebreaker session on Saturday 7 March (~1.5 hours)
Available every Saturday from 14 March - 13 June 2026 (~7.5 hours per session)
Nice-to-have
Technical AI alignment research experience
Previous experience running technical workshops or bootcamps
Experience with distributed/remote teaching
Why join us?
Shape APAC AI safety talent development: Help expand the first dedicated technical AI safety program across the Asia-Pacific region
Teaching autonomy: Help define how we teach AI safety concepts
Career development: Deepen your technical and teaching abilities
Community impact: Build technical AI safety communities across multiple cities
Collaborative team: Work alongside other TAs
Application Process
Applications close Friday 23 January 2026. We aim to:
Begin reviewing and interviewing candidates as applications come in
If we find the right people before January 23 we might hire them. Early applications are encouraged!
Conduct interviews between 2 January and 19 February
Make final decisions by Friday 21 February
Questions? Contact yanni@taraprogram.org and zac@taraprogram.org
