Agastya Goel

Agastya Goel

Intern, Applied AI @ Ramp

About

I love thinking about computer science and physics and finding ways to apply them to my everyday life. I also enjoy performing through choir and theater, as well as playing tennis and board games.

Country

United States

City

Stanford

Industry

Computer Software

Skill

Large Language Models (LLM), Physics, Mathematics, Data Structures, Algorithms, Algorithm Analysis, Machine Learning Algorithms, HTML, JavaScript, Java, C++

Experience

Ramp

Intern, Applied AI

Ramp

LinkedIn
2025-6 - 2025-9 · 4 mos
Autoscience Institute

SWE Intern

Autoscience Institute

LinkedIn
2025-2 - 2025-2 · 1 mo
Research Science Institute (RSI)

Researcher

Research Science Institute (RSI)

LinkedIn
2024-6 - 2024-8 · 3 mos

Cambridge, Massachusetts, United States

I replicated human survey results by replacing human survey participants with large language models (LLMs) answering on behalf of generated personas, in order to better understand both human and LLM behavior. I introduced a key innovation: an extra modeling step after receiving LLM responses. This approach yields significantly improved predictions compared to a uniform guess, and confirms previously discovered biases in LLMs. Paper in progress..

Euler Circle (Palo Alto)

Researcher

Euler Circle (Palo Alto)

2021-6 - 2024-2 · 2 yrs 9 mos

RS-complete Cycle Types, 2022 PUBLICATION: The Australasian Journal of Combinatorics - Published PAPER: [2301.07216] RSK-Complete Cycle Decompositions https://arxiv.org/abs/2301.07216 Understanding the specific cycle decompositions achieving RSK-completeness is foundational in various mathematical areas, facilitating algorithmic improvements, combinatorial enumeration, theoretical advancements, and providing connections to representation theory and beyond. In this paper, we characterize the class of cycle decompositions that can achieve all Young tableau shapes (except the trivial ones with a single row or single column) under the Robinson–Schensted–Knuth (RSK) correspondence, a property that we call RSK-completeness. We prove that for even n, cyclic permutations comprise the only fixed cycle decomposition that is RSK-complete. For odd n, cyclic permutations and almost cyclic permutations which have a cycle of length n−1 are the only RSK-complete cycle decompositions. Invited talk at CombinaTexas 2024, Texas A&M

PRIMES-USA

Researcher

PRIMES-USA

2023-1 - 2023-11 · 11 mos

Arithmetic of Semisubtractive Semidomains, 2023 (Group Project, PRIMES-USA) PUBLICATION: Journal of Algebra and its Applications - Published PAPER: [2311.07060] Arithmetic of semisubtractive semidomains https://arxiv.org/abs/2311.07060 Factorization theory studies the decomposition of elements into irreducible factors within various algebraic structures. This investigation allows us to prove general results about algebraic objects based on just a few properties. A primary objective in this field is to gauge the extent to which an object deviates from having unique factorizations, and this paper almost completely answers this question with regard to semisubtractive semidomains. In this paper, we study the arithmetic of semisubtractive semidomains (i.e., semidomains S for which either s ∈ S or − s ∈ S for every s ∈ 𝒢(S)). Specifically, we provide necessary and sufficient conditions for a semisubtractive semidomain to be atomic, to satisfy the ascending chain condition on principals ideals, to be a bounded factorization semidomain, and to be a finite factorization semidomain, which subsequent relaxations of the property of having unique factorizations. This gives a near-complete characterization of the ascension and descension of these properties between integral domains and semisubtractive semidomains.

Education

Massachusetts Institute of Technology

Massachusetts Institute of Technology

LinkedIn

Computer Science

Gunn High School

Gunn High School

LinkedIn
Stanford Online High School

Stanford Online High School

LinkedIn

Agastya Goel's Contact Information

Email

******@***.com

Phone

(**) *** ****

Find the Right Leads
Find Verified Contact Data

Try with: Jensen Huang @ nvidia.com Click to autofill
LeadContact awards, five-star ratings, and GDPR compliance badges

What LeadContact does well

Find verified emails, phone numbers, and decision-makers with 98% accuracy.

Find Leads

Find Leads

Find the right people by company, role, industry, location, and more.

925M+ professional profiles

Find Leads
Find Emails

Find Emails

Access verified email addresses for your target contacts.

657M+ emails

Find Emails
Find Phone Numbers

Find Phone Numbers

Get cross-validated phone data from multiple top sources.

239M+ phone numbers

Find Phone Numbers

More Accurate. Lower Cost.

Find contact data in 1 tool with 98% accuracy

LeadContact integrates leading enrichment tools to deliver more accurate contact data—without paying for each one.

LeadContact Logo
Competitor Tools

All these = $289 per month

Great conversations start with the right contact.

It’s time to find yours.