Anjor Kanekar
I'm a computational physicist turned Forward Deployed Engineer. I spent 7 years at Palantir building and hiring FDE teams, and I write here about how that role actually works — and about doing real physics research again, this time with AI.
About Me
PhD in Computational Physics from the University of Maryland. I spent 7 years at Palantir as a Forward Deployed Engineer and hiring manager, conducted over 1,000 technical interviews, and trained hundreds of interviewers. Featured in The Pragmatic Engineer on how FDE organizations work.
I run Platypus Technologies, where I advise companies on how to run forward deployed engineering, and train their FDEs. The framework behind that work, the Platypus FDE Blueprint, is published free.
I've also gone back to plasma physics research, handing more of the work to AI each time. I rebuilt my PhD-era code with Claude. The code and a paper describing it are on GitHub. One study looked like a result until a closer check showed it was already known, so I closed it. My current study leaves the method to the model. More on the research page.
Writing
Forward Deployed Engineering
The role is having a moment, and most of what's written about it is muddled. My running attempt to pin it down lives on the FDE page. It covers whether you need the role, what it is, the fork AI has opened up, and how to grow and train FDEs.
- How Do You Grow FDEs?: Most FDEs have to be grown. Define the role first, then the training.
- The Fast Layer and the Slow Layer: The tools change every few months. The craft of making a deployment work barely changes.
- If LEGO Had Forward Deployed Engineers — What actually makes an FDE different from a really good consultant
- The FDE Fork: Platform or Outcomes — The strategic choice founders don't realise they're making
- Is FDE a Bubble? — The title is spreading faster than the reasoning behind it
- The FDE Manifesto — What the role is for, and why most companies get it wrong
AI + Physics Research
- The Boundaries of Delegation: What I hand over to AI in research, one year on
- A Week in Vienna — What I learned returning to a plasma-physics working meeting after twelve years
- The Autonomy Gradient — What AI can and can't do in physics research
- Writing a Physics Paper with Claude — Workflow, iterations, and hallucinations caught in review
Hiring & Organizations
- The Unreasonable Effectiveness of Hiring Assholes — Why conviction and ego are separate traits
- Context Matters: Hiring is Not Fungible — Why generic hiring processes fail
- Comand AI: A Hiring Case Study — How we designed hiring for mission-driven engineers
Work With Me
- Get new FDEs ready to lead deployments → The FDE Programme. A two-day exercise on a realistic client case, then six weeks of coaching on their own deployments from people who have done the job. Each FDE gets a written verdict on whether they're ready.
- Set up or fix an FDE function → Platypus Technologies. Start with the two-week assessment: a written readout of where you sit against the Blueprint, what is wrong with the plan, and what to change in the next ninety days.
- Speak to your team or portfolio → the talks page. Panels, roundtables, and sessions for leadership teams and investors.