Skip to content
View wasim's full-sized avatar
7 followers · 7 following

Organizations

@k-not

Block or report wasim

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
wasim/README.md

Wasim Akram

Principal Data & AI Engineer @ VML MAP (WPP) · P1 Affiliated Researcher @ AI Centre Denmark · Copenhagen 🇩🇰

I build production-grade agentic AI systems and resilient data/ML platforms, while studying how sparsity and modularity emerge in dense language models—and how to turn that structure into real efficiency wins.

Right now I'm working through Antikythera's What Is Intelligence? curriculum and replicating Anthropic's interpretability studies to stay grounded before pushing the frontier.

  • 🔬 Current project: Scaling Specialization in Dense LMs (sparsity & modularity vs. model size; dynamic-k inference)
  • 🧠 Focus: sparse autoencoders (SAEs), mechanistic interpretability, scaling laws, evaluation/reliability
  • 🛠️ Stack: Python, HF/Transformers, Snowflake/BigQuery, GCP/AWS, dbt, Spark/Glue, Vertex/Cloud Run, CI/CD, cost/perf

Links: Website · GitHub · LinkedIn · AI Centre DK · VML MAP · Email: mailwasim [at] gmail dot com

Stay hungry. Stay foolish.

Pinned Loading

  1. learning-terraform learning-terraform Public

    Forked from LinkedInLearning/learning-terraform-3087701

    This repo is for the Linkedin Learning course: Learning Terraform

    HCL

  2. nanochat nanochat Public

    Forked from karpathy/nanochat

    The best nano ChatGPT from wasim

    Python

  3. scaling-specialization-dense-lms scaling-specialization-dense-lms Public

    Do dense LMs develop MoE-like specialization as they scale? Measure it, visualize it, and turn it into speed.

    Python 1