Ayush Goyal

Associate Consultant

Los Angeles, California, United States2 yrs 3 mos experience

Key Highlights

Strong foundation in mathematics and computer science.
Experience in machine learning and quantitative finance.
Proven ability to enhance system efficiency and interpretability.

Stackforce AI infers this person is a Fintech and AI research specialist with a focus on machine learning and quantitative analysis.

Contact

ayushgoyal5720@gmail.com LinkedIn

Skills

Core Skills

Machine LearningQuantitative Analysis (finance)Natural Language Processing (nlp)XaiComputer Vision

Other Skills

AlgorithmsAutoCADC++Data StructuresDependency ParserGenerative Adversarial Networks (GANs)Image SegmentationJavaLarge Language Models (LLM)PyTorchPython (Programming Language)QuantizationResearchRetrieval-Augmented Generation (RAG)Teamwork

About

I’m a Computer Science graduate student at USC (graduating December 2025) with a strong foundation in mathematics from IIT Delhi. My journey has taken me through quantitative finance, machine learning research, and large-scale AI systems — experiences that have shaped how I approach problems: with rigor, creativity, and a focus on impact. Over the years, I’ve had the chance to work across very different domains: from developing trading algorithms and predictive models in finance, to advancing explainable AI and model compression in research, to building multimodal systems with social applications. Across all of them, I’ve found a common thread: I thrive on making complex systems more efficient, interpretable, and useful. My interests sit at the intersection of machine learning, quantitative finance, and high-performance engineering. I’m motivated by environments that are fast-paced, collaborative, and intellectually challenging — where I can push boundaries while building solutions that matter. 🔎 Currently seeking full-time opportunities starting January 2026 where I can bring together my background in ML research, quantitative modeling, and systems engineering to solve high-impact problems.

Experience

Goldman sachs

Quant Strategist Intern (Summer Analyst)

Jun 2025 – Aug 2025 · 2 mos · New York, New York, United States · On-site

On the Mortgage Strats team, I engineered features and trained a Random Forest (scikit-learn) model to predict Commercial Mortgage Backed Securities (CMBS) building valuations. On the IRP SMM Strats team, I extended the Spreader algorithm in Java to improve performance under aggressive market conditions and investigated execution slippage against arrival market prices, identifying opportunities worth approximately +$137K per $1mm duration traded.

Machine LearningQuantitative Analysis (Finance)JavaPython (Programming Language)Algorithms

University of southern california

Graduate Research Assistant

Aug 2024 – Jan 2025 · 5 mos · Los Angeles, California, United States · On-site

Developed question-answer copilots for semiconductor fabrication using zero-shot multimodal RAG, in collaboration with Prof. Rehan Kapadia and Prof. Swabha Swayamdipta. The project integrated human-training conversations to enhance accuracy in low-resource domains.

Natural Language Processing (NLP)Retrieval-Augmented Generation (RAG)Large Language Models (LLM)Python (Programming Language)

Korea advanced institute of science and technology

Research Assistant

Aug 2023 – Jun 2024 · 10 mos · Daejeon, South Korea · Remote

Worked at the Statistical Artificial Intelligence Lab (SAIL), KAIST under the guidance of Prof. Jaesik Choi, where I enhanced explainability efficacy by leveraging dependency parser-based syntactic units, achieving a 7% increase in metric scores. I evaluated methods such as LIME, SHAP, IG, LRP, and Attention on text classification models for the Plug and Play XAI project, and developed a pipeline to identify the most suitable explainer algorithm for specific text-based models, thereby improving their trustworthiness and reliability.

XAILarge Language Models (LLM)Dependency ParserPython (Programming Language)

Adobe

Research Intern

Jun 2022 – Aug 2022 · 2 mos · Remote | Bangalore, India

Worked on model compression for GANs, where I analyzed the layer-wise sensitivity of the StyleGAN architecture by estimating the average Hessian matrix trace using the Hutchinson algorithm. I implemented fixed-precision and mixed-precision Quantization-Aware Training techniques in PyTorch to optimize both model size and performance, ultimately achieving a compression ratio of about 6.5× with negligible degradation in FID scores compared to the full-precision model.

Machine LearningResearchGenerative Adversarial Networks (GANs)QuantizationPyTorch

Mathematics society, iit delhi

2 roles

Coordinator

Aug 2021 – May 2022 · 9 mos · Delhi, India

Executive

Dec 2020 – Aug 2021 · 8 mos · Delhi, India

University of sussex

Research Intern

May 2021 – Jul 2021 · 2 mos · Remote | Falmer, Brighton, East Sussex, England

Weakly Supervised Semantic Segmentation
Investigated the feature of equivariance of image segmentation to affine transformations, to achieve near fully supervised semantic segmentation based on image-level annotations
Explored the paper and code of self-supervised equivariant attention mechanism (SEAM), exploiting the above feature

ResearchImage SegmentationComputer Vision

Board for student welfare (bsw), iit delhi

Academic Mentor

Nov 2020 – Feb 2021 · 3 mos

Responsible for guiding freshers to focus on their academic and holistic development. I assisted them in academics by holding various doubt clearing sessions and help them adapt to a totally different type of academic environment.