Project summary

I recently learned that there is a mechinterp conference in Boston on Friday August 22nd 2025.

Unfortunately the registration and submission deadline has already passed, so I don't have the option to present at the venue. However I'd still find it valuable to participate in discussions and network.

By default, I cannot plan on going, unless someone decides to fly me out for the weekend.

What are this project's goals? How will you achieve them?

I have been interviewing for mid-to-senior level research/software engineering positions in industry at AI safety orgs during the past ~2 weeks. Attending the conference would support this immediate job search in various ways (e.g by exploring potential opportunities at Goodfire that would be a good fit for my skillset), and also help set up spontaneous connections for future collaborations.

How will this funding be used?

There are two options:

Bus: $800 ($700 for transportation costs, $0 for lodging, $100 for miscellaneous expenses).
Plane: $2000 ($1000 for transportation costs, $800 for lodging, $200 for miscellaneous expenses).

Who is on your team? What's your track record on similar projects?

I'm only asking for travel funding for myself. I'm an independent alignment researcher based in Toronto. I have mentored student projects in the past via SPAR, but I'm still learning the ropes of publishing, so I've only been on the author list for a handful of workshop papers, including one at CVPR on interpreting large text to image diffusion models which recently lead to my first citation.

However, I wasn't first author on that paper so it wasn't a primary focus at the time, and honestly I'd rate my performance on that particular project as poor. I'm better much positioned to contribute to interpretability projects now that I have more directly relevant experience. I also have a strong engineering background in areas adjacent to ML, such as security and compliance at Microsoft Canada.

What are the most likely causes and outcomes if this project fails?

Causes:

I'm fairly confident that the funding goal is unrealistic on this timeline and that the value add is unclear. I'm quite flexible and capable of adjusting my plans on short notice, but I believe that most potential funders would prefer to invest in other projects that are tied to more legible and predictable outcomes.

Outcomes:

I'd be happy to connect with the participants next time there's an opportunity to do so in-person (such as at NeurIPs or BlackboxNLP). If there's a significant overlap in our interests, then the conversations would require lower bandwith, and could be done over fully remote channels such as text/call.

How much money have you raised in the last 12 months, and from where?

I haven't raised money in the last 12 months.