Saanvi

Aspiring Applied AI & Machine Learning Student's Portfolio

About Me

Hi! I'm Saanvi, an aspiring computational researcher developing robust, uncertainty-aware machine learning models for medical imaging and diagnostic medicine with experience in multimodal diagnostic architectures, domain adaptation, and synthetic genetic circuit design. I'm also a figure skater, percussionist, and fiction writer who believes in bringing creative solutions to all my research projects.

Education & Coursework

Wayzata High School

Plymouth, MN โ€” Class of 2029

Premier public high school recognized for academic excellence and STEM rigor.

MIT Beaver Works Summer Institute (BWSI)

Virtual

Applied Engineering & AI (Fall 2025) | Computational Problem Solving (Spring 2026)

Online & University Computer Science Coursework

Academic Honors

Independent Research: ColoScreen

ColoScreen - Machine Learning for Early Triage of Visible Rectal Bleeding

ColoScreen is a medical AI research project designed leveraging Generative AI (Stable Diffusion) and Bayesian inference to detect GI pathologies i.e. detect and classify blood on toilet paper using computer vision. By leveraging a hybrid pipeline of physical data acquisition, digital compositing, and generative synthetic data (500K images), the model accounts for variances in substrate texture, lighting, and blood morphology to provide accurate triage.

Current Phase: Data Engineering & Model Architecture Design

Research & Technical Projects

Deep Learning for Medical Image Diagnostics

July 2025 โ€“ July 2026

Lead Developer โ€” UMN Data Science Initiative / Independent | Python, PyTorch, Torchvision, Optuna

CampQuest: AI-Powered Enrichment Discovery Platform

June 2025 โ€“ Present

Lead Developer | Python, Supabase/SQL, Llama 3

UCSC Python and Research (PYaR) & ReComBio

Spring 2026

Student Researcher โ€” Virtual | R, Python, Anaconda, Jupyter Notebooks

Summer Research Experience

Stanford AIMI High School Summer Research Internship

Stanford, CA (Virtual) | June 2026

Summer Research Intern โ€” 1st Place Research Competition (Perfect Methodology Score)

  • Awarded 1st Place (7 teams) with perfect methodology score from five AIMI faculty judges for PneumonAI pneumonia classifier; led interpretability and uncertainty quantification.
  • Designed multimodal DenseNet-121 with Albumentations, metadata fusion and two-stage training; ablation studies contributed to top model (Swin-Tiny AUC 0.862) and multimodal validation accuracy gains (from approximately 85% โ†’ 90-93%).
  • Implemented Grad-CAM visualizations to identify failure modes (textual artifacts, view-position labels); evaluated Bayesian CNN with Monte Carlo Dropout.

MIT Jameel Clinic AI & Health Summer Bootcamp

Cambridge, MA (In-Person) | July 2026

Bootcamp Participant โ€” 1st Place Hackathon Winner

  • Led 6-person team to 1st Place (10 teams) on 25,000-record synthetic EHR for EpiPen prediction with temporal cutoff at Visit 2; evaluated Random Forest, Logistic Regression, XGBoost, and MLP via ROC-AUC, F1, Recall, and Brier score.
  • Built end-to-end preprocessing pipeline including one-hot encoding, missing value imputation, and 80:20 train-val split with 5-fold CV; serialized production models with joblib for downstream inference; invited as 2027 TA.

Stanford iGEM Bioengineering Research Program (SiBRP)

Genetic Circuit | July 2026

Lead Designer & Primary Author | Selected from 5,000+ Applicants

  • Led design of a dual-input AND-gate genetic circuit in E. coli Nissle 1917 using split-T7 RNA polymerase to simultaneously detect nitric oxide (PnorV) and ROS (PsoxS) as early gut proxies for Parkinson's disease.
  • Integrated an sfGFP fluorescent reporter module, designed auxotrophic biocontainment mechanisms, and evaluated potential clinical applications.

Stanford SCIยณ Mini-Fellowship | Hamline Microscopy Camp

Stanford, CA (Virtual) / St. Paul, MN | June 2025 โ€“ July 2025

Advanced Molecular Imaging Techniques & ESEM Experiments

  • SCIยณ: Advanced molecular imaging for in vivo diagnostics.
  • Hamline: Conducted ESEM analysis of tubular structures in hair samples from twins and compared with dog and cat hairs.

ML - Police Search Bias Prediction

Research Scholar| UMN Basic Machine Leaning Camp | Python, Scikit-learn, Pandas, Matplotlib

  • Machine learning model analyzing potential bias in police searches. Combines exploratory data analysis, fairness metrics, and model comparison to predict search likelihood.

Leadership & Outreach

CodeUnity

Co-Founder | Launched January 26, 2026 | MIT BWSI Official Sponsor of Hackathon

Co-founded nonprofit teaching free CS to underrepresented middle-school students; grew to 70+ students across 7 states and 3 countries with MIT BWSI sponsorship.

Mission: Making computer science accessible by teaching "Coding through Art & Apps" to middle schoolers.

Curriculum: TurtleLabs โ€“ combining creative visual output with foundational programming concepts & C.O.R.E teaching app building using Python.

CodeUnity bridges the gap between artistic expression and computational thinking, helping young students discover that coding can be both powerful and playful. Through interactive Python Turtle projects, students learn loops, functions, and problem-solving while creating visual artwork they're proud to share.

Enrichment Programs

  • University of Minnesota Advanced AI/ML & Python Camps.
  • Fermi National Accelerator Laboratory: Saturday Morning Physics.
  • Kode With Klossy Data Science Camp.
  • Columbia University SEM Program.

Extracurriculars

Competitive Programming & Mathematics

Active Competitor | Wayzata High School (2025 โ€“ Present)

  • USACO Bronze Division: 2nd Place Intermediate at CodeHER International Competition (an USACO-styled coding competition with participants from 25+ countries in partnership with the Competitive Programming Initiative [CPI] and sponsored by NYU Tandon, Jane Street, and Non Trivial, 2026); Top 20% in the MIT Informatics Tournament Open Division. Varsity Math team competitor.

Performing Arts & Athletics

Percussionist, Dancer, & Competitor | Wayzata High School & Community (2021 โ€“ Present)

  • Music & Dance: Percussion (Marimba, Snare Drum, Timpani) โ€” GTCYS Concert Orchestra, MacPhail Rimshots: 6-member auditioned percussion ensemble, UMN High School Honor Band, 2026 MBDA State Honor Band. Kathak Dancer: Advanced-level Indian classical dance at Katha Dance Theatre (Lucknow Gharana, Guru Rita Mustaphi).
  • Competitive Athletics: Figure Skating: 2000+ hours trained, competed in 117 events with 82 top-three finishes, 4x ISI World Gold, USFSA Pre-Gold skater, Junior Company member American Ice Theatre (AIT); USA Fencing Nationals Qualifier.

Humanities & Creative Writing

Iowa Young Writers' Studio & Varsity Speech Competitor | Online / Wayzata High School (2025 โ€“ 2026)

  • Selected and attended Iowa Young Writers' Studio online fiction writing program. Drafting a novel, The Threads of Diocian. Two Scholastic Silver Keys. Creative Expression Speech finalist in 4 competitions.

Technical Skills

Programming Languages:Python, Java, C++, Wolfram (Mathematica), R
Frameworks & Libraries:NumPy, Pandas, Matplotlib, Scikit-learn, BeautifulSoup, Albumentations, Tkinter, Flask, PyTorch, Torchvision, OpenCV
Databases:MySQL, Supabase, PostgreSQL
Platforms:VSCode, Jupyter, Replit, GitHub, Google Colab, Kali Linux
Web Technologies:HTML, CSS, JavaScript
Tools:Tableau, Git, Bash, Pixlr, LaTeX, Canva

GitHub

Explore my open-source work across data science, AI, and software engineering.

ColoScreen Repository

Detailing the workflow, image capture and synthetic image generation using Stable Diffusion (Juggernaut XL)/ LoRA training in ComfyUI and Albumentations.

Machine Learning Repository

AI/ML notebooks using scikit-learn and PyTorch. Covers data preprocessing, model evaluation, visualization, and interpretation.

Java Coding Repository

Collection of Java programs, algorithms, and USACO prep problems demonstrating OOP, data structures, and algorithmic practice.

Python Coding Repository

Foundational Python exercises and projects โ€” scripts, games, and learning notebooks showcasing growing technical fluency.

GitHub Profile & Stats

ColoScreen Repository

ColoScreen Last Commit ColoScreen Commit Activity ColoScreen Stars

Machine Learning Repository

Machine Learning Last Commit Machine Learning Commit Activity Machine Learning Stars

Python Coding Repository

Python Coding Last Commit Python Coding Commit Activity Python Coding Stars

Java Coding Repository

Java Coding Last Commit Java Coding Commit Activity Java Coding Stars
GitHub Stats Top Languages
Profile Views

Certificates & Badges

MIT BWSI

MIT BWSI Yes! You can do Computational Problem Solving

MIT BWSI

MIT BWSI Girls Who Can Learn Applied Engineering, AI & Future Technologies

CS50P

EdX- Harvard University CS50's Introduction to Programming with Python

U-Mich

University of Michigan โ€” Joy of Coding

UMN

University of Minnesota โ€” Advanced AI Camp

Fermilab

Fermi National Accelerator Laboratory

AlphaStar

AlphaStar Intermediate Java

Kode With Klossy

Kode With Klossy โ€” Data Science

Wolfram

Wolfram Summer Program

MnRI

UMN - Minnesota Robotics Institute

Affiliations

AoPS

AoPS

MIT BWSI

MIT BWSI

UMN

University of Minnesota

KWK

Kode With Klossy

Wolfram

Wolfram

cyberarts.io

cyberarts.io

edX CS50P

edX CS50P

Fermilab

Fermi National Accelerator Laboratory

Mehta+

Mehta+

Hennepin

Hennepin Technical College

Contact

๐Ÿ“ง Email: saanvi.research@gmail.com

๐Ÿ’ป GitHub: github.com/SalmaLilad

ยฉ 2026 Saanvi | Coding with Curiosity & Creativity