SOFTWARE ENGINEER
Tarun Vaidhyanathan
Large-scale systems. Clean logic. Real impact.
CS & Economics at Stony Brook — building at the intersection of finance and technology.
ABOUT
A Bit About Me
I'm currently a senior at Stony Brook University, majoring in Computer Science and Economics. I want to learn and experience as much as I can, both in tech and in the financial world. Those two fields feed off each other in ways that keep me genuinely excited to show up every day.
What I enjoy most is building large systems. My strength has always been in the logic of constructing them, understanding how pieces fit together and being able to implement what feels intuitive. I love working with cutting-edge technology, whether that's building an ML compiler that lowers PyTorch models into hardware-optimized executables at the COMPAS Lab, LLM-powered and large-scale data pipelines at Amazon, or optimizing no-touch trading on in-memory frameworks at Piper Sandler.
When I'm not building, I'm usually cooking and eating, working out, or hanging out with my friends!
Stony Brook University
B.S. Computer Science (Honors)
B.A. Economics
Expected May 2027
New York Area
NYC preferred · Open to relocation
Relevant Coursework
Operating Systems, Cloud Computing, Networks, Machine Learning, Data Science
Hobbies
EXPERIENCE
Where I've Worked
Amazon
New York, NY
Software Development Engineer Intern – Percolate
Jun 2026 – Aug 2026
- •Built reusable Airflow/EMR/Spark pipelines populating a separate Glue/Athena analytics layer, enabling observability and conversational analytics while reducing projected storage costs by $35.6K/month
- •Automated dataset onboarding with AWS CDK, reducing provisioning to schema/config/DAG changes
- •Developed a Spring Boot MCP server with 17 Scala-based tools automating end-to-end regression testing for Spark/Hadoop changes on jobs processing 100s of TB/day, projected to save $36K/year in developer time
Software Development Engineer Intern – Shopping Guides
Sep 2025 – Nov 2025
- •Built LLM keyword-generation and validation pipelines with Step Functions, S3, Lambda, Claude 3.7 on Bedrock, and GPT-OSS for 300K+ Shopping Guide keywords from high-intent queries
- •Built large-scale Scala pipelines over 100M+ search queries, measuring 99.65% revenue coverage and supporting changes that drove a 105% revenue increase
Stony Brook, NY
Undergraduate Researcher
May 2026 – Present
- •Building a model-agnostic second-gen MLISA compiler that captures PyTorch graphs and lowers them through a multipass graph-to-C pipeline into hardware-optimized executables while preserving loop/function hierarchy
- •Porting the optimized first-gen MLISA operator/runtime stack, whose measured GPU inference matched TorchInductor across four NVIDIA GPUs, into the new compiler
Piper Sandler
Greenwich, CT
Equities Trading Technology Intern
Jun 2025 – Aug 2025
- •Designed latency-testing infrastructure for GigaSpaces in-memory data grids, benchmarking no-touch equities workflows and reducing on-call investigation time by ~20%
- •Integrated Datadog APIs and migrated legacy Perl scripts to PowerShell, reducing system latency by 20% and improving metrics accuracy by 12%
PROJECTS
What I've Built
Weather Derivatives Trading System
Streaming LightGBM probabilistic model that combines live weather observations with NBM forecasts to predict daily-high temperature outcomes. Powers automated trading bots for Polymarket US and Kalshi with an automated order management system. Services are deployed with Docker and backed by a persistent SQLite store.
Fourier Fund Analytics Suite
Portfolio analytics suite for 45+ analysts spanning efficient-frontier optimization, correlation analysis, backtesting, volatility, and asset allocation, with a backtested Markowitz mean-variance optimizer.
CPA Client Portal
Secure ASP.NET Core client portal for a CPA firm with role-based access, AES-256-GCM encryption, Azure Blob Storage, and document-integrity checks.
Seawolf Accessibility
Next.js/Python app for accessible campus routing with C-based OpenStreetMap parsing, Dijkstra's algorithm, and indoor mapping. Improved routing with scikit-learn regression and KNN, raising accuracy by 23% and creating alternative routes.
Reactive Collision UAV
Authored SDF/XML PX4 Gazebo simulations to automate 20+ flight and collision tests. Developed UAV stress-testing tools with OpenCV/MATLAB to measure deformations under 9.8N loads.
SKILLS
Technical Expertise
Languages
Frameworks & Libraries
Cloud & Data
Focus Areas
LEADERSHIP
Beyond the Code
Website Development Club
President
Jan 2024 – Present
Scaled membership from 60 to 140+ via hackathons and industry-speaker professional development events. Led biweekly workshops on LLM wrappers, ML visualization, and webapp development.
Fourier Fund
Financial Data Scientist
Aug 2024 – Present
Stony Brook Student Led Investment Fund. Developed the Portfolio Optimizer and participated in stock voting and market recaps. Presented the Optimizer's usage, projected asset allocation, and analyst tool suite to 40+ fund members.
CONTACT
Let's Connect
Interested in fintech, distributed systems, or just want to chat? Feel free to reach out.