Ridwan Alam

Blogs

Musings on backends, data, and whatever else is on my mind.

  1. Do LLMs Really Understand Math? A Knowledge Space Theory Perspective

    An LLM can solve the hard problem and fail the arithmetic underneath it, and accuracy still scores that as a win. An ETH Zurich paper uses Knowledge Space Theory to test whether a model's right answers respect their prerequisites, and finds knowledge that is fragmented rather than hierarchical.

    • AI Engineering
    • Gen AI
    • LLM Evaluation
  2. Distributed Rate Limiting

    A counter in process memory stops meaning anything the moment you run a second replica: each instance enforces its own limit, so ten pods let through ten times the traffic you configured. What it costs to keep one shared count honest.

    • System Design
    • Distributed Systems
  3. Message Brokers and Event Streaming with Kafka

    Kafka is not a queue. It is a distributed, append-only commit log that many consumers read at their own pace, and every feature it has follows from that one fact: partitions, consumer groups, offsets, delivery semantics, retention and the outbox trick.

    • Backend Engineering
    • System Design
    • Distributed Systems
  4. Speculative Decoding: Making LLMs 3x Faster with Zero Quality Loss

    A small, fast draft model proposes the next few tokens and the large one verifies them in a single batched pass. The speedup is 2-3x, and the output distribution is provably identical to running the large model alone.

    • AI Engineering
    • Gen AI
    • LLM Inference
  5. Chaos Monkey: Breaking Your Own System in Production on Purpose

    A SIGCOMM paper argues the recovery code in a network controller is only ever tested by breaking the live network on purpose: what to break, how to prove the break is survivable first, and why the bug can live entirely in the milliseconds before the system settles.

    • Backend Engineering
    • System Design
    • Distributed Systems