MS CS @ NYU + Stern FinTech • Ranked #1/2,800 in programming • McKinsey Forward Alum

I taught self-driving cars to see at 80% autonomy. Now I teach algorithms to see money.

Went from scoring 98.1% accuracy grading 8K papers with BERT → building VaR dashboards that cut risk insights by 40% → optimizing sub-2s latency trading systems. My Monte Carlo simulations run 10,000+ scenarios before your coffee gets cold.

Current Arsenal:

• Event-driven trading strategies with CVaR and Sortino ratios
• Real-time risk monitoring systems (10K+ concurrent users, 95% uptime)
• Production ML: GLIDE models at 96% precision, pix2pix GANs at 90% accuracy
• Financial forecasting models improving $10M+ datasets by 12% (McKinsey/AWS)

Microsoft Research: Broke and rebuilt generative models—improved fidelity 25% over 80K training steps

McKinsey: Built AWS Lambda + DynamoDB forecasting systems serving enterprise SMEs

NYU Stern: Shipped iOS meal analysis app with Flask + MongoDB backend, 90%+ accuracy in <2s

Blackbox: PyMC3 Bayesian models hitting 80.83% prediction accuracy with 15% error reduction

Note: Stats reflect public repos only. The real action happens in production where milliseconds cost millions and "good enough" is career suicide.

I don't do leetcode. I do systems where Python meets P&L, algorithms have attitude, and the only thing moving faster than the code is the money it manages.

Seeking: Teams where technical debt excites me more than it scares me.

📧 utsavd7@gmail.com • 🔗 linkedin.com/in/utsavd7 • 🌐 utsavdoshi.vercel.app

From SD-WANs to cGANs, from autonomous vehicles to autonomous trading—if it computes, I've probably optimized it.