Hi there! I’m Aniket, a self-taught Machine Learning Engineer with a background in Computer Science & Engineering.
Currently, I’m working on open-source projects at Lightning AI as a Developer Advocate.

You can find most of my new blogs on LLMs, Distributed Training, and Machine Learning on Lightning AI’s blog.

I enjoy building and deploying machine learning systems at scale. In the past, I’ve created computer vision and NLP-based systems that automated significant portions of business operations for a major e-commerce company.
I also enjoy tinkering with open-source tools and have both created and currently maintain several open-source projects, including Gradsflow and Lit-GPT.

DevRel Experience

I joined Lightning AI (formerly PyTorch Lightning) after being a community contributor and core maintainer of Lightning Flash. I made significant contributions from the library’s launch day and hosted meetups and workshops to educate developers.

I have been working on building the community, creating technical content, and using my engineering skills to build flagship products. Some of my achievements here are:

  • Grew the London community by 10x and hosted multiple in-person community meetups.
  • Created high-quality content that drove the majority of the traffic to the Lightning.ai website. My blog was trending on HackerNews as well.
  • Live-streamed about LLMs and distributed model training on our brand Discord channel, growing the community to more than 1K.
  • Managed Twitter and LinkedIn for posting technical content and driving engagements with the developer community.

Engineering Experience

Before a DevRel, I am a Machine Learning engineer. I have worked with startups to build and deploy machine learning systems in production. I successfully scaled ML services to handle millions of requests per hour by implementing a messaging queue, automatic batching, and various other optimizations. My open-source project, Chitra, is actively used in production environments.

I spearheaded the development of Lightning AI’s flagship product, Muse. I was responsible for deploying stable diffusion and scaling it to support 1K+ concurrent users.

Recently, my work has focused on distributed training of LLMs and deployment. I contribute to Lit-GPT, an open-source LLM training and finetuning library developed by Lightning AI. Additionally, I integrated the LLM evaluation framework LM-Eval Harness and HELM (Stanford’s evaluation framework) to enhance the evaluation of LLMs.