⬡ DSK-OS
Home
Research
Projects
Experience
Education
Skills
Awards
Contact
Beyond
AVAILABLE
QUANT · RESEARCH
SATX
00:00:00
◆ARBA // 17–29% Lower Regret
◆SSRN // 4,813 Views
◆UT Health // Student Researcher
◆WKL World Champion // Black Belt
◆Latin Oratory // #2 National
◆SAT // 1580 · 800 Math
◆National Merit Commended
◆MIMIC-IV // 21.5% Regret Reduction
◆ARBA // 17–29% Lower Regret
◆SSRN // 4,813 Views
◆UT Health // Student Researcher
◆WKL World Champion // Black Belt
◆Latin Oratory // #2 National
◆SAT // 1580 · 800 Math
◆National Merit Commended
◆MIMIC-IV // 21.5% Regret Reduction
Identity dsk@dsk-os
Darshan Sathish Kumar — Quantitative Research, UTSA Honors Mathematics and Computer Science
Darshan
Sathish Kumar
Mathematics · Quantitative Research · Computational Modeling
UTSA Honors undergraduate studying Mathematics and Computer Science. I work on problems at the intersection of quantitative finance, optimization, statistical modeling, and computational systems — from empirical asset-pricing research to adaptive hospital-resource allocation.
My work includes independent research in momentum and market microstructure, robust optimization for healthcare operations, systematic trading infrastructure, and biomedical research at UT Health San Antonio.
Python · R
SSRN Published
UTSA Honors
Class of 2029
Select Research Area 4 areas
ARBA
Adaptive-Regret Bed Allocation
Optimization · Healthcare Ops
Momentum
Large-Cap Equities 12–1 Strategy
Asset Pricing · Quant Finance
UT Health
Liver Disease & Metabolic Syndrome
Biomedical · Statistical Modeling
Microstructure
LOB Imbalance & Return Predictability
Market Microstructure · Kraken
Adaptive-Regret Bed Allocation CUSUM-triggered robust optimization
A CUSUM-Triggered Robust Optimization Framework for Non-Stationary Healthcare Environments
Method Poisson CUSUM + MDP + MILP
Sim 20-bed ED · SimPy
Regret O(log T)
Solver HiGHS oracle
| Scenario | Static Robust | Adaptive | Reduction |
|---|---|---|---|
| Baseline diurnal | 31,013 | 21,903 | 29.4% · p<.001 |
| Flash flood 3× surge | 53,607 | 44,205 | 17.5% · p<.001 |
| Creeping crisis drift | 46,978 | 37,738 | 19.7% · p<.001 |
| Acuity flip | 42,613 | 30,694 | 28.0% · p<.001 |
Key Findings
Detection is fast: CUSUM flags step shifts in 1.13h (bound 3.17h), drifts in 4.83h — zero false alarms at baseline.
Validates on real data: 21.5% regret reduction on a 236-visit MIMIC-IV clinical cohort (p < 0.001).
Bounded regret: O(log T) vs linear O(T) for static robust — re-centers the uncertainty set on detection.
Selected Projects 5 systems
| # | Project | Stack | Status |
|---|---|---|---|
| 01 | Systematic Trading Framework Research, execution & risk analytics infrastructure | Python · Alpaca | ● Active |
| 02 | NetGuard Real-time intrusion detection platform | FastAPI · React | ● Completed |
| 03 | ER Flow Dashboard Patient-flow bottleneck detection system | pandas · matplotlib | ● Active |
| 04 | Order Flow & Microstructure Kraken LOB imbalance · 10k+ ticks | Python · LOB | ● Research |
| 05 | Adaptive-Regret Bed Allocation CUSUM + robust optimization · SimPy / MIMIC-IV | Python · SimPy | ◆ Research Project |
Systematic Trading Framework
Six strategies (weekly reversal, cross-sectional reversal, cointegration pairs, low-vol tilt, covered-call overlay, composite) with automated Alpaca paper execution and historical backtesting. Automated reporting: equity curve, drawdown, rolling Sharpe, SPY alpha/beta/R², per-order PnL, monthly returns. Reproducible trade logs.
↗ github.com/dshan12/systematic-trading-frameworkER Flow Dashboard
Modeled a real bottleneck witnessed as a hospital volunteer. Event-driven pipeline processing 300+ timestamped patient-flow logs. Rebuilt around boarding-time after direct nurse feedback. Detects bottlenecks ~3 min earlier than manual observation on synthetic datasets.
↗ github.com/dshan12/ER-DashboardAdaptive-Regret Bed Allocation ★ primary home: research.exe
Poisson CUSUM change detector coupled to a gamma-robust bed reservation policy with O(log T) regret. SimPy 20-bed ED sim: 17–29% lower regret across 4 surge regimes (all p < 0.001), 21.5% on MIMIC-IV validation. 71 tests, HiGHS hindsight oracle.
↗ github.com/dshan12/ARBANetGuard — Real-Time Intrusion Detection
Self-hosted IDS: Scapy sniffer → Redis pub/sub → 5-signature rules engine + 3-model ML ensemble (Isolation Forest, autoencoder, K-Means) → Postgres → FastAPI (REST + WebSocket) → React dashboard. Docker Compose one-command demo.
↗ github.com/dshan12/netguardOrder Flow & Market Microstructure
Kraken BTC/USD (5,768 ticks) + ETH/USD (5,002 ticks): top-of-book imbalance, Newey-West predictive regressions, 85s BTC signal half-life, velocity +8–23% ΔR². Event-driven market-making sim included.
↗ github.com/dshan12/Market-MicrostructureResearch
UT Health San Antonio
Jun 2025 – Present
Student Researcher — Liver Disease & Metabolic Syndrome · Dr. Mengwei Zang
- Promoted from Student Research Assistant after one year; lead statistical modeling, longitudinal analysis, and data-pipeline development
- Quantitative analysis of longitudinal data from 100+ mice across preclinical MASLD cohorts; evaluated three investigational compounds (Bempedoic Acid, ACLY inhibitor 78C, NMN) using GraphPad Prism and Excel
- Presented findings directly to UT Health Dean and faculty
Selective Programs
SIG High School Discovery Program
May 2026
Quantitative Trading / Market Making
- Trader/researcher-led probability and market-making problems; decision-making under uncertainty
RBFCU Youth Ambassador
Jun 2026
Finance & Banking Program
- Selected 1 of 20 from 120+ applicants; intensive finance program with industry professionals
Healthcare
University Health / Methodist Healthcare
Clinical Training
Sterile Processing Technician — CBSPD Certified
- Four-month clinical training program across two hospital systems; BLS (AHA), HIPAA, and OSHA certified
Methodist Hospital
Jan 2025 – Present
Student Volunteer — Clinical Units (200+ hours)
- Assisted nurses across 6 clinical units; took vitals and supported care for 60+ patients
- Direct observations motivated the ER Flow Dashboard project
Service
Blood Cancer United
Service
Hall of Fame Team Member — Fundraising Campaign
- Three-month campaign; team raised $60,000+ for blood-cancer research across Texas; individual Hall of Fame recognition
Smithsonian Institution
Mar 2022 – Present
Transcription Volunteer — Remote (150+ hours)
- Transcribed and digitized historical documents to improve public access to Smithsonian collections
Simulations
JPMorgan Chase
May 2025
Quantitative Research Simulation — Forage
- Forecast natural-gas prices with time-series modeling; priced a storage contract; credit-risk model with dynamic-programming FICO bucketing
Bank of America
Simulation
Global Markets Sales & Trading Simulation
- Analyzed rates, sector performance, and volatility; produced a client-facing market-insight report
Academic Profile
UTSA Honors College
San Antonio, TX · Mathematics + Computer Science · Fall 2026 – Spring 2029
Fall 2026 — Current Courses
◎
MAT 1313 — Algebra and Number Systems
◎
MAT 2213 — Calculus III
◎
MAT 2253 — Applied Linear Algebra
◎
PHY 1943 — Physics for Scientists and Engineers I
◎
PHY 1951 — Physics for Scientists and Engineers I Lab
◎
AIS 1203 — Academic Inquiry (Honors)
◎
HON 1000 — Honors 101
Dual Enrollment
✦
Northwest Vista College — Dual Enrollment · 45 college credits · GPA 3.91
✦
Relevant Coursework — Calculus I–II
High School
Northside Health Careers High School
San Antonio, TX · Medical Magnet · Aug 2022 – May 2026 · Rank 55/221
GPA / 100
98.5
SAT Total
1580
Math SAT
800
Self-Directed Coursework
✦
Harvard CS50 — Introduction to Computer Science · Completed
✦
MITx 18.02 — Multivariable Calculus: Vectors & Derivatives · Verified Track · Completed
✦
Wharton / Coursera — Introduction to Corporate Finance · Completed
◎
MIT OCW — Financial Mathematics · In Progress
School had no CS, business, or finance courses until senior year. All advanced quantitative skills built independently. Exhausted all available math by junior year.
Languages
PythonRSQLTypeScript
Used in: ARBA · Momentum · NetGuard · Trading
Quantitative Research
Factor ModelsNewey-West HACBacktestingMarket MicrostructureTime Series
Used in: Momentum · Microstructure · Trading
Optimization & Simulation
Robust OptimizationMILPDynamic ProgrammingSimPyHiGHS
Used in: ARBA · ER Dashboard · Trading
ML & Data
scikit-learnpandasNumPystatsmodels
Used in: NetGuard · Momentum · UT Health
Systems & Web
FastAPIReactPostgreSQLRedisDockerGitHub ActionsScapy
Used in: NetGuard · Trading
Languages (Natural)
English — FluentTamil — FluentHindi — Conversational
Competition
WKL World Champion — Traditional Forms2025
WKL National Runner-Up — Point Sparring2025
2nd Place Nationals — Latin Oratory, National JCL2024
1st Place State — Latin Oratory, Texas JCL2025
HOSA First Aid — 1st Local · Top 10 Texas2023–24
3rd Place Nationals — Greek Derivatives, NatJCL2023
Academic
National Merit Commended Scholar2025
AP Scholar with Distinction2025
SAT 1580 · 800 Math—
Research Recognition
Single-Author SSRN Preprint — Momentum2026
4,813 Views · 385 Downloads—
Contact
Let's build
something remarkable.
something remarkable.
Graduating May 2029. Looking for quant research, systematic trading, and quant dev internships at the intersection of quantitative methods and real-world systems.
✉darshansathishkumar@gmail.comEmail ⌥github.com/dshan12GitHub ◈SSRN Author ProfileResearch ◎linkedin.com/in/darshansathishkumarLinkedIn ↓Darshan_Sathish_Kumar_Resume.pdfResumeQuick Facts
| Status | Available · Grad May 2029 |
| Location | San Antonio, TX |
| Education | UTSA Honors · Math + CS '29 |
| Focus | Quant Research · Trading |
| Fun Fact | 🥋 World Karate Champion |
Beyond Research
🥋 Karate
Since 2012
Black Belt · WKL World Champion · Junior Instructor
- Traditional forms world champion; point-sparring national finalist; mentor to younger and lower-belt students
🏛 Classics
Since 2022
Latin Oratory · NJCL Praetor · Category Leader
- 2nd nationally, 1st in Texas; elected chapter Praetor; peer tutor for calculus & Python
🩺 Healthcare
Clinical
Hospital Experience · HOSA · BLS Certified
- 200+ volunteer hours across clinical units; sterile-processing training; HOSA Top 10 Texas
🤝 Service
Community
Blood Cancer United · Smithsonian
- Hall of Fame fundraiser ($60K+ team campaign); 150+ hours transcribing Smithsonian collections
👤
Home
📊
Research
⚙️
Projects
🏢
Experience
🎓
Education
💻
Skills
🏆
Awards
✉️
Contact
🌿
Beyond