Skip to content
View Jeevesh28's full-sized avatar
7 followers · 1 following
  • University of Wisconsin–Madison
  • Madison, Wisconsin
  • LinkedIn in/jeevesh28

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Block or report Jeevesh28

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 is supported. This note will only be visible to you.
Report abuse

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

Report abuse
Jeevesh28/README.md

Hey there, I am Jeevesh! 👋

MS Computer Science at UW–Madison, graduating May 2027. I work on LLMs, ML systems, and the engineering underneath them. Most recently a Data Science Intern at Cloudflare. Before graduate school, 2+ years at Oracle building backend services and data infrastructure.

Looking for full-time SDE and ML engineering roles starting Summer 2027.

Selected work

Open source: RSEM (merged upstream) Modernized a widely cited RNA-Seq quantification tool from C++98 to C++17 and removed its legacy Boost dependencies, with a drop-in RNG so existing analyses still reproduce bit-for-bit. Added a gold-standard test suite across four aligners with cross-platform CI, then ported the Perl drivers to Python 3. #233 · #234 · #235

Intelligent-Bistro A mobile ordering app where guests manage a cart by talking to an assistant. Claude parses natural-language orders into validated tool calls. UI taps and AI calls share one cart-mutation path, so the model's view never drifts from the user's, and pricing is always computed server-side. Expo, React Native, Express, TypeScript.

CoT-Faithfulness A behavioral and mechanistic audit of chain-of-thought faithfulness across Llama 3.1 8B, Gemma 3 12B, and Claude Opus 4.6. Four CoT perturbations re-injected via assistant prefilling, plus attention analysis. Faithfulness falls as capability rises.

Certified-Temporal-GNN Certified robustness for temporal graph neural networks. Replaces the per-window Bonferroni union bound with one shared edge-flip realization under a single Clopper–Pearson bound: sound under Neyman–Pearson, strictly tighter, and up to +41.7 points certified accuracy on real Elliptic Bitcoin data.

SynthShield Turns sensitive tabular data into synthetic clones with formal privacy guarantees. T5-small fine-tuned with LoRA under DP-SGD, reaching ε≈4. FastAPI and Next.js on Cloud Run, built in 24 hours at CheeseHacks 2026.

Background

Experience

  • Cloudflare, Data Science Intern, Jun 2026 – Aug 2026
  • UW–Madison Data Science Institute, Open Source Developer, Feb 2026 – Jun 2026
  • Oracle, Applications Developer, Jun 2023 – Aug 2025

Education

  • UW–Madison, MS Computer Science, 2025–2027
  • IIT Kharagpur, B.Tech (Hons.) Mechanical Engineering, Minor in Mathematics & Computing, 2019–2023

Connect

Pinned Loading

  1. deweylab/RSEM deweylab/RSEM Public

    RSEM: accurate quantification of gene and isoform expression from RNA-Seq data

    C++ 477 123

  2. CoT-Faithfulness CoT-Faithfulness Public

    Behavioral and mechanistic audit of chain-of-thought faithfulness across Llama 3.1, Gemma 3 and Claude Opus, using perturbation re-injection and attention analysis.

    Jupyter Notebook 1

  3. Certified-Temporal-GNN Certified-Temporal-GNN Public

    Forked from rohankrgupta/certified-temporal-gnn

    Certified robustness for temporal GNNs via shared-noise randomized smoothing. One shared edge-flip realization under a single Clopper-Pearson bound, strictly tighter than per-window Bonferroni.

    Jupyter Notebook

  4. SynthShield SynthShield Public

    Forked from saksham-2000/SynthShield

    Differentially private synthetic tabular data. T5-small with LoRA under DP-SGD, served through FastAPI and Next.js on Cloud Run.

    TypeScript

  5. Intelligent-Bistro Intelligent-Bistro Public

    Mobile ordering app where a conversational AI assistant and the UI share one validated cart-mutation path, so the model's view of the cart never drifts from the user's.

    TypeScript

  6. Facility-Location-Problem Facility-Location-Problem Public

    Uncapacitated and capacitated facility location problems solved with GurobiPy, including UFLP under disruption scenarios.

    Jupyter Notebook 10