work.ipynb
Python 3 | idle

Curriculum Vitae

full resume on Overleaf → view PDF

In [ ]:
import cv
cv.social_links()

§ Skills

In [ ]:
cv.skills()
Out[ ]:
categorytechnologies
Languages Python · Java · C · C++ · SQL · TypeScript · JavaScript · HTML/CSS · R
Software Tools Amazon Web Services · Google Cloud Platform · Databricks · Docker · Claude Code · Codex · Gemini · TensorFlow · PyTorch · Pandas · NumPy · HuggingFace · Apache Spark · Git/GitHub · Linux Systems
Domains Cloud Development · AI/ML · Generative AI · Internet of Things · Systems Engineering · Data Structures & Algorithms · Full-Stack Development
Data Viz Grafana · Tableau · Matplotlib · Plotly · D3.js
Web React · Astro · Solid.js · Node.js · FastAPI

§ Work Experience

In [ ]:
# 7 roles · sorted by date desc
cv.work_experience()
Out[ ]:
  • Productionalized a serverless data pipeline and query service that aggregates hardware metadata and deployment events from 4 upstream sources into a unified REST API, replacing fragmented investigations for HWEng teams.
  • Designed a novel ingestion pipeline with Annapurna Labs firmware telemetry, making visible 32% of bad deployment incidents across manufacturing lines, reducing average incident resolution time from 13 days to under 1 day, and enabling early detection of incompatible firmware installations.
  • Designed and deployed an unsupervised machine learning API using SageMaker, Lambda, and S3 to analyze 10 million+ deployments across hardware rollouts and automatically surface anomalous behavior and unsafe deployments.
  • Replaced manual signal inspection with automated anomaly scoring and deployment-level recommendations, reducing rollback investigation time by 75% and enabling engineers to evaluate rollout health 4x faster.
  • Operationalized an ML-based sustainability classification pipeline that converted 10,000+ subjective data points into standardized scores used in internal evaluations and customer-facing dashboards.
  • Worked directly with the CEO to translate qualitative sustainability criteria into reproducible data-processing and classification workflows, improving consistency across company analyses.
  • Supported 800+ students across DSC 40A, DSC 100, COGS 9 (Kyle Shannon), and COGS 108 (Jason Fleischer) through office hours, grading, and technical instruction in Python, SQL, databases, algorithms, and machine learning.
  • Served as software lead by developing and maintaining course websites and migrating automated grading infrastructure to a new platform via custom scripting, increasing grading speed by 50% across large-enrollment courses.
  • DSC 40A Website →
  • COGS 9 Website →
  • Developed user-interface designs using HTML, CSS, and Canva for a start-up finance app.
  • Beta-tested iOS app using TestFlight and provided detailed feedback to the developer team.
  • Managed the Sports Jam summer programs for middle and elementary schoolers, using sports and crafts to instill values of open communication and teamwork.
  • Led a team of high school students working with park-rangers to restore California Nursery Historic Park for its re-opening as part of the Youth Service Corps (YSC).
  • Promoted Simplify's AI-powered job application platform to 100+ students at UC San Diego.

§ Research

In [ ]:
cv.research()
Out[ ]:
  • Collaborating with Ph.D. student Akshay Nagarajan on an IoT-based educational platform leveraging embedded systems to improve K–12 learning outcomes.
  • Responsible for constructing the backend infrastructure, database design, and data engineering pipelines.
  • Contributing to a forthcoming research publication based on this work.

§ Leadership

In [ ]:
cv.leadership()
Out[ ]:
  • Leading one of the largest data science student organizations in the U.S. (500+ members) by driving strategy, operations, and cross-campus engagement through industry talks, academic panels, technical workshops, hackathons, and projects.
  • Previously as Software Director (2024–2025), coordinated the redesign of the DS3 website using React & GitHub to enhance UX for 500+ members, sponsors, and partners.
  • DS3 Website →

§ Featured Projects

In [ ]:
from ansh.portfolio import projects
# 16 published · sorted by date desc
cv.projects()
Out[ ]:
projecttagsdescription
PlantDex Machine Learning · Computer Vision · Python ML-powered plant identification and classification app built for DS3
DS3 Website Full-Stack Development · TypeScript · React Main website for the Data Science Student Society at UC San Diego
DataHacks 2026 Full-Stack Development · TypeScript Official website for DataHacks 2026 — UCSD's data science hackathon
Smart AC System IoT · Embedded Systems · Systems Programming Cyber-physical smart air conditioning system with Arduino and live sensor data
Twitch Recommender Machine Learning · Recommender Systems · Python Content recommendation system for Twitch streams using collaborative filtering
LukaBot Machine Learning · Reinforcement Learning · Python Reinforcement Learning agent learning how to shoot a basketball
DataSurg Data Visualization · JavaScript A Visual Exploration of Korean Surgical Data
HeatMappers Machine Learning · Data Analysis · Python National Data Platform Fire-Ready Forests Data Challenge
AutoBots Machine Learning · Reinforcement Learning · Python Using Reinforcement Learning for autonomous vehicles with Carla & DonkeySim
BikeWatching Data Visualization · JavaScript Interactive visualization of bike traffic and station activity across San Diego
Recipes-n-Ratings Machine Learning · Python · Data Analysis Do Higher-Calorie Recipes recieve Higher Ratings on food.com?
Personal Website & Home Foresting Full-Stack Development An online space to digitally introduce myself & write about my rare-plant hobby
Fantasy Basketball Assistant Full-Stack Development · APIs · Python Software to assist with NBA Fantasy Sports
Criminal Justice AI Machine Learning · Data Analysis · Python AI model to to remove human bias from sentencing & improve COMPAS
SongSpace Full-Stack Development · Python · Databases Prototype-version of Spotify
WebCrawler Java Created a Web Crawler prototype from scratch in Java