a.k
/
about me

about me

currently

doing ai stuff @ lossfunk โ†’. I love working on funky problems, and the curiosity driven nature of lossfunk helps me explore to my fullest.

One of my main reasons for joining lossfunk was to build interpretable models; which not only helps humans, but also explain their refusal of answers.


previously

Undergraduate Researcher @ UIUC (advised by Prof. Darko Marinov) โ†—

  • Engineered a workflow using LLMs to perform semantic evaluation on software mutants, building a process to classify code changes as equivalent to guide subsequent manual analysis.
  • Developed an analysis script to identify "fishy" mutants by replicating and extending the heuristics from the BugFarm ICSE paper.

SDE Intern @ Google

  • Implemented a data-driven approach with the help of XAI methods and LIME technique to refine inappropriate question detection, leading to a 20% reduction in false positives based on user feedback.
  • Worked with chain-of-thought prompting, specifically a template-based approach, to enhance the reasoning capabilities of large language models for multi-step tasks; Improved the F1 score by 24% compared to baseline models without CoT prompting.
  • Developed a solution utilizing symbolic reasoning for questions, enabling efficient handling of complex mathematical problems like Lorentz ODE proofs.

occasionally + shoutout

If I am not doomscrolling on Twitter or Insta, I spend time watching YouTube documentaries (Minecraft ARGs are my current favorite). I also read papers or go biking with my friends because I love to explore places. I'm always grateful to my family and my friends because they have been with me through the ups and downs of my life <3