Software Engineer III
Sep 2026 – Present
- Leading development of a new full-stack agentic platform in FactSet's newly
formed AI business unit — enabling clients to build their own AI agents, or
deploy FactSet-defined ones, and run them directly on FactSet's proprietary
financial data
- Designed and implemented the platform's core capabilities — an MCP-enabled
architecture with agent orchestration, agentic workflows, prebuilt scheduled
routines, and a dashboard surfacing the results of every agent run
Software Engineer II
Sep 2025 – Sep 2026
- Drove the ground-up rewrite of FactSet's flagship next-generation web
platform, owning the redesign and implementation end-to-end (Vue, TypeScript)
and embedding the FactSet AI assistant into a completely reimagined workflow —
launched to 150,000+ daily active users with over
99% client satisfaction
- Selected by leadership as one of 6 nominees across
the entire business unit for the Dealmakers & Wealth Trailblazer Award —
recognizing ownership and leadership of the next-generation platform
rewrite
Software Engineer I
Mar 2024 – Sep 2025
- Handpicked for the CEO-led task force that rebuilt FactSet's iOS and Android
apps — owned the Markets tab, delivering market data and charting across all
asset classes to 50,000+ users in a well-received
relaunch
- Engineered the tab as shared Vue + TypeScript webviews embedded in native
Swift and Kotlin shells — one codebase serving both platforms
- Maintained and upgraded core web platform infrastructure through nearly
2× user growth — optimizing for millisecond-level
performance and workflow efficiency across
150,000+ daily users
Agentic AI
FinTech
Vue
TypeScript
iOS & Android
- Subject matter expert across freelance AI training-data platforms including Alignerr
— leading agentic coding projects that produce high-quality human data for training
and evaluating frontier large language models
- Partner directly with frontier AI labs to research and design novel project concepts
that advance their models' agentic coding capabilities by targeting specific, observed
failure modes
- Run projects end-to-end — translating client model-training specifications into
precise annotation guidelines and owning data quality and delivery across the full
labeling pipeline
- Research new quality-assurance methodologies to guarantee label quality and
instruction adherence — from rubric design and gold-standard calibration tasks to
inter-annotator agreement analysis
Agentic Coding
RLHF
LLM Post-Training
Data Quality
- Unlocked fully programmatic portfolio construction for buy-side quantitative clients —
optimal portfolios built directly from client-defined constraints and objectives, with
zero dependency on the user interface
- Architected and developed 12 C++ REST API endpoints —
3 for each of FactSet's 4
in-house and vendor portfolio optimization engines
- Constructed Jupyter Notebooks demonstrating end-to-end API automation workflows to
buy-side quant clients
C++
Portfolio Optimization
REST APIs
Jupyter
- Reconstructed and maintained the official website for the UMich Engineering Career
Fair — hosted by Tau Beta Pi and the Society of Women Engineers — drawing
26,000+ views
- Revamped the sponsorship portal that brought in $263,100
in revenue from 260+ companies
- Oversaw the Website Committee Chairs, providing insight to the fair's directors to
inform decisions
- Organized 5,000+ students and
400+ corporate recruiters during the two-day event
HTML & CSS
JavaScript
Web Hosting
- Developed a 3,400-line test suite in Vector CAPL, run in
CANoe, to verify and debug the automation tools used to test automotive hardware —
increasing testing efficiency by 80%
- Created and executed 200+ software and systems test cases
on physical ECUs (electronic control units) bound for automotive OEM products, using
Vector CANalyzer to simulate CAN bus traffic
- Generated and submitted test reports with defect lists directly to customers and the
global engineering team
Automotive ECUs
Vector CAPL
CANoe & CANalyzer
CAN Bus