Talk to Arpit about
- databases
- distributed systems
- soft skills
How Arpit thinks
Judgment over automation
Values human judgment, communication and context more than AI‑generated output.
For you: Focus on sharpening your judgment and communication to stay valuable.
Agent orchestration like k8s
Sees agent orchestration as analogous to Kubernetes, emphasizing delegation and retry handling.
For you: When designing agent systems, think of clear delegation and retry boundaries.
Sharpen selective edges
Prioritizes improving chosen abilities while allowing others to wane.
For you: Identify which skills matter most to you and double‑down on them.
Ask this first
- How do you balance AI‑generated code with human judgment in your projects?
- What challenges did you face building px0.ai’s LSP integration?
- Can you share insights on scaling databases at Razorpay?
Arpit's brain
Recently updated
GitHub PR workflow: running `px0 <github-pr-url>` opens a PR directly in px0 and shows the full comment thread.
In late September 2026 I posted a series of X replies that covered my thoughts on cognitive decline, the limits of AI‑generated code, the…
Px0Performance.startup_time = less than 1 s (sometimes claimed 1 ms)
I recapped Java’s G1 garbage collector, noting its region‑based heap layout and pause‑budget strategy.
I described Google Zanzibar’s Leopard Index, which stores two pre‑computed lookup tables—member → group and group → group—to make…
I observed that Google’s AX orchestrator works like “kubernetes for agent executions”, handling delegation and synthesis of sub‑agents…
I announced the TrueForge/TrueFoundry hackathon on 26 Sept 2026, a free in‑person event with a Rs 3,00,000 prize pool and 350‑400 seats.
I maintain a résumé‑style list of projects spanning 2011‑2021, primarily done at IIITH, TCS Innovation Lab, COEP, Unacademy and Practo.
I am a senior software engineer and engineering leader with over 12 years of experience at Google, Amazon, Unacademy and Razorpay. I hold…