Neev Parikh
This stream focuses on empirical projects that improve our ability to evaluate model capabilities or enable us to understand or evaluate model monitorability. An ideal project culminates in a research output (conference/Arxiv paper or research blogpost with artifacts).
Stream overview
Some examples of projects I'd be excited about are below. These project ideas are less well-scoped by me but I'd be interested and excited if scholars had clear ideas of what to work on here:
- Investigate how to build monitorability / control evaluations with natural model goals or side tasks.
- Leverage model-based environment simulations to build better, more realistic
- Investigate swarm dynamics including if such interactions could increase monitor collusion.
- Investigate the "motivated reasoning" present in 2026 AI incidents
- Build better alignment evaluations based on real incidents.
- Build better red-team attack policies for monitorability evaluations.
- Models are increasingly trained with synthetically generated data. Does this mean that models will be able to instinctively "underperform" (or sandbag) by behaving like the less-capable model that generated their training data without needing to reason about this?
- Try and understand how various recently proposed or discussed neuralese/latent reasoning architectures (like Astra or current open-source default reasoning paradigm) and estimate how the alignment tax for visible-chain-of-thought has changed over time. This might look like compiling the presented results and analyzing trends but also might involve collecting new data by benchmarking the relevant models.
Light-touch (.5 hours of weekly 1:1s)
SF Bay Area
Indifferent
Indifferent
Mentors
Neev likes to make computers do interesting things, deeply understand concepts and build interesting, useful tools. He is currently thinking about AI alignment, control, and evaluations, and work with frontier models at METR.
Recent work done involves MALT, training models to fool monitors in QA settings and RE-Bench. Neev previously worked at Stripe and CSM, and did a concurrent BSc/MSc in Computer Science at Brown.
Fellows we are looking for
An ideal fellow has a strong AI research (software engineering is a plus) background. It's important that they are self-motivated and can make weekly progress with little intervention. If you are interested in working on non-concretely scoped projects, I would expect mentees to have the ability to write well-scoped project proposals, with realistic planned milestones and deliverables. Evidence of successful projects here would be very helpful in evaluating this. A fellow can be a PhD student and they can work on a paper that will be part of their thesis.
- Past research experience in the relevant domain with public artifacts that can be discussed or reviewed
- Evidence of exceptional speed of iteration and ability to derisk ideas quickly.
Project selection
I will talk through project ideas with the scholar