Resume

Partnership With Children

Data Analytics & Engineering Intern · Jan 2025 – Jan 2026 · NYC

• Engineered a FERPA-compliant automated ETL pipeline in Python, JavaScript, and Bash to migrate protected student records from Apricot into SharePoint, reducing a month-long manual workflow to 7 minutes while maintaining data integrity and real-time access.
• Developed a secure Azure-hosted Streamlit dashboard with Docker and Microsoft SSO that gave leadership reliable visibility into program performance, informing grant renewals, staffing allocation, and program expansion across NYC schools, supported by VBA automation and GitHub Actions CI/CD.
• Consolidated 10+ years of NYSED, InfoHub, and internal program datasets using advanced SQL joins and unions, validating and reconciling discrepancies to within 0.5% margin of error and standardizing metrics supporting grant reporting and district evaluations across 30,000+ students in New York City.
• Migrated the entire Apricot case-management platform — 10+ years of FERPA- and HIPAA-sensitive records — into Salesforce across 32 NYC schools, partnering with non-technical stakeholders to design SOQL-driven notification Flows that fit their actual workflows, and established data governance standards adopted organization-wide.

Python JavaScript SQL Bash Azure Docker Streamlit Salesforce

Neuromatch

NeuroAI · Computational Neuroscience · Deep Learning Researcher · Summers 2024 – 2026

• 2026 (NeuroAI): Implemented Node Perturbation learning on a multilayer perceptron for Project TardyGrad, benchmarking biologically plausible learning against Backprop and Hebbian learning on MNIST classification; NP achieved 99% test accuracy on par with Backprop in the ANN setting, with degraded performance under spiking neuron conditions identifying a meaningful boundary for gradient-free learning.
• 2025 (Computational Neuroscience): Fit Generalized Linear Models to mouse visual cortex recordings spanning V1 and higher visual areas, decoding behavioral outcomes including lick and reward events and encoding neural responses to visual stimuli across learning stages, extending findings from Zhong et al. (Nature, 2025); co-authored an Impact Scholars proposal designing a temporal-difference learning extension targeting reward anticipation and prediction error signals in anterior HVAs.
• 2024 (Deep Learning): Built end-to-end pipelines in Python and Pandas to ingest, join, and transform 1.5M+ COCO-MS image captions; fine-tuned and validated RoBERTa, BERT, and BLIP with PyTorch and HuggingFace, applying MLR and cosine similarity metrics to extract object-category sentiment patterns; results contributed to a peer-reviewed publication (arXiv:2409.09560).

Python PyTorch HuggingFace Pandas GLMs

Annalect

Data Engineering Apprentice · Aug 2022 – Nov 2022

• Optimized large-scale Airflow-orchestrated pipelines for pharma and media clients by restructuring Jinja2 templates, modifying existing DAGs, and refactoring event-driven AWS Lambda trigger logic, reducing data latency by 30%.
• Contributed to ETL/ELT workflow automation across AWS (Lambda, EMR, S3, Redshift, Glue) and PostgreSQL, leveraging PySpark and Redshift Spectrum to process terabyte-scale data, supporting pipelines driving a 10% CPA drop and 12% performance lift across client pipelines.

Python SQL Apache Airflow AWS PySpark PostgreSQL

The Knowledge House

Technical Instructor · Aug 2025 – Present

• Designed and facilitated hands-on technical instruction for teen and adult learners across Python, Jupyter, introductory data science, HTML/CSS/JavaScript, Git/GitHub, generative AI, prompt engineering, and cybersecurity, translating complex concepts into accessible, project-based lessons grounded in inquiry, productive struggle, and real-world applications.
• Developed and delivered responsible AI and digital literacy curricula on AI bias, cultural representation, privacy, hallucinations, prompt design, phishing detection, and secure technology use; coached learners through technical projects while collaborating with curriculum and instructional leadership to strengthen classroom engagement, learner confidence, and skill.

Python JavaScript Git/GitHub Asana Figma Jupyter Generative AI Cybersecurity

Data Science Fellow · Sep 2021 – Sep 2022

• Built ETL pipelines in Python and Bash to ingest and clean 500+ alumni records, integrated Google API for user authentication and data ingestion, and conducted exploratory data analysis, including geospatial engagement patterns with Pandas, supporting a 20% increase in platform user engagement in a cross-functional team environment.

Python SQL Bash HTML/CSS Pandas

Real-Time Equity Analytics Platform

Kafka, Databricks, Snowflake, dbt · Oct 2025 – Dec 2025

• Designed and deployed a production-grade batch and streaming ELT medallion pipeline ingesting NYC School Climate and CDC Social Vulnerability Index data via Kafka and Dataproc into GCS Bronze/Silver layers, with Databricks Spark Structured Streaming and dbt-modeled Snowflake Gold marts powering district equity dashboards for policy and resource decisions.
• Built a LangChain and Chroma RAG service with a FastAPI backend and React/TypeScript UI, embedding 5K+ SVI and school climate documents to enable natural language querying of equity metrics over validated Snowflake Gold data.
• Provisioned the entire cloud infrastructure from scratch using Terraform across GCP, Snowflake, and Databricks, implementing DevSecOps practices: detect-secrets scanning, least-privilege IAM roles, and segregated GitHub Actions CI/CD with zero credentials in version control.

Python SQL Kafka Databricks Snowflake dbt Terraform

Automated Compliance-Grade Document ETL Pipeline

Python, JavaScript, Bash · Mar 2025 – May 2025

• Built a FERPA- and HIPAA-compliant ETL pipeline to extract, validate, and migrate 10,000+ protected student records from Apricot into Excel and SharePoint using Puppeteer-clustered PDF automation and a Tkinter GUI, reducing manual processing from months to 7 minutes.

Python JavaScript Bash

Regression Analysis on SARS-CoV-2 RNA Concentrations in NYC Wastewater

Data Science · Python, SQL · Oct 2023 – Oct 2024

• Built a production-aware RESTful API and custom SQLiteDB class with decorator-based error handling, timer logging, and input validation decorators to ingest and store NYC OpenData SARS-CoV-2 wastewater concentrations, implementing an optimized unique key extraction algorithm with O(n·m_max) time complexity and O(k) space complexity for efficient JSON schema discovery across 5,754 records.
• Applied GridSearchCV-optimized A/B testing with 5-fold cross-validation across five regression models (SGD, Polynomial, Lasso, and Random Forest) across all 14 NYC wastewater facility sites, achieving R² up to 0.95 and identifying consistent seasonal concentration peaks aligned with citywide COVID-19 surge patterns to support public health reporting.

Python SQL scikit-learn REST API SQLite

Student Adaptability Level Classifier in Online Education

Data Science · Python · Jul 2024

• Implemented a complete supervised classification pipeline on a Bangladeshi online education dataset from hypothesis through EDA, One-Hot Encoding, and feature engineering to Naïve Bayes, KNN, and Random Forest classifiers evaluated via confusion matrices, cross-validation, and A/B testing across Low, Moderate, and High adaptability levels; Random Forest achieved the highest average F1 score of 0.837.
• Addressed significant class imbalance across adaptability levels using RandomOverSampler and RandomUnderSampler strategies, then selected the final model via A/B testing across 100 cross-validation iterations; Random Forest's F1 of 0.837 outperformed KNN (0.803) and Naïve Bayes (0.685) across all three classes.

Python scikit-learn Pandas Matplotlib Seaborn