Machine Learning & Quant Software Engineer

Victor Roberto Esquivel Estrada

Mexico City ******@******.** +527********* Software Engineer (2013) at ITESM code91.mx/vree in/esquivelvictor github.com/VicEsquivel SKILLS (Proficient highlighted)

MISCELLANEOUS: TensorflowJS • MongoDB • SQL • Electron • Google Firebase • Expo • ML BACKTESTING/BROKER:Alpaca • Interactive Brokers • Nautilus Trader • Backtesting.py • Backtrader.py BACKEND: C# • Express • NodeJS • Java • Python

FRONTEND: ReactJS • React Native • JavaScript • Bootstrap • HTML • CSS • ChartJS • JQuery 3D RELATED:Unity 3D/2D C# • Shader CG • Blender 3D modeling • VR • AR • C++ SOFT SKILLS: Problem solver • Results oriented • Logical thinker • Competitive • Self motivated LANGUAGES:English 100% • Spanish 100%

EXPERIENCE

COMPANY: Quant Evolution

ROLE:Co-Founder and Technical Lead

PERIOD:October 2025 - Present

MAIN GOAL:

Quant Evolution is a quantitative research and financial technology company focused on developing machine-learning and evolutionary-computation systems for financial-market analysis and algorithmic trading.

As Co-Founder, I define the company’s technical and research direction and lead the development of its quantitative-trading infrastructure.

Key responsibilities and projects:

-Limit Order Book Prediction System: Designed transformer models that process high-frequency LOB data and generate Buy, Sell, and Idle signals from prices, sizes, counts, spread, imbalance, and multiple book levels.

-Custom Machine-Learning Architectures: Built specialized TensorFlow.js components, including fixed-lag and last-query attention, positional embeddings, convolutional context layers, tokenizers, and labeling systems.

-Feature Engineering and Representation Iteration: Continuously test features, tokenization, normalization, labels, and data structures to find the richest and most learnable representation for transformer models.

-Transformer-Based Signal Research: Adapt transformers to LOB data using a custom discrete token representation that supports signal classification for trading strategies.

-Distributed Training Infrastructure: Built a multi-computer training system using Node.js, TCP/IP, TensorFlow.js, packed gradients, and synchronized model weights.

-Algorithmic Scalping and Backtesting: Developed backtesting systems that model spreads, bid/ask costs, take-profit, stop-loss, confidence thresholds, holding periods, and concurrent positions.

-Broker and Trading-Platform Integration: Integrated experimental workflows with NautilusTrader and designed connectivity for Alpaca and Interactive Brokers.

-Evolutionary Trading Research: Lead research in genetic algorithms, genetic programming, symbolic regression, evolved decision trees, feature selection, portfolio optimization, and multi-objective evaluation.

-Probabilistic Market Modeling: Researched HMMs, MCMC, KDE, and regime detection for forecasting price movements and identifying market states.

-Technical and Strategic Leadership: Define research priorities, evaluate results, guide the technical roadmap, and identify scalable quantitative-trading opportunities. COMPANY: Code 91 S.A. de C.V.

ROLE:Founder and CEO

PERIOD:November 2014 - Present

MAIN PROJECTS:

Main projects developed along Code 91’s history:

-Genetic Algorithm Server: Using TensorflowJS, NodeJS, Express.

-Genetic Programming Module: I developed a module using a genetic algorithm that evolves a population of decision trees to discover the optimal dynamic structure. This structure can be trained on custom features and labels for classification or regression tasks.

-Scalping: I created and optimized a stock market scalping script. Alpaca and IBrokers API.

-OrderBook simulator: I built a custom simulator that replicates a stock market limit order book, supporting order placement, matching, cancellation, and real-time updates for strategy testing. It can load data of official Limit Order Books, such as LOBSTER Project Name: GTD APP • Client: Grupo Bimbo • México • Web, Android, iOS • +14 months

-Link-1

-Is an internal project management app for their HQ.

-It has around 200 daily active users today, and has been in production environment since 2020.

-I started and finished the project (Management, Commercial, Technical execution): 1.Identifying the requirements

2.Created a multi-screen visual UI/UX mock-up model in photoshop, HTML, JS 3.Iterated on the mock-ups, and once they were signed, I started the development (frontend in ReactJs for web, and React Native for the Android and iOS app using Expo). 4.I used Google Firebase for the backend.

Project Name: VIRay • Client: Virtual Interior Systems GmbH • Germany • Unity 3D • ~30 months

-Link-1 • Link-2 • Link-3 • Link-4

-This is a Unity 3D photorealistic virtual tours software (Initially only for kitchen market, but then after my involvement, it evolved to also be able to support real-estate projects)

-My role was very extensive. When I entered the project VIRay was in version 116. After my work we reached version 337. And almost tripled the features VIRay had and expanded the business vision.

-Coded a runtime OBJ importer for VIRay and polished the DAE file importer.

-Created multiple shaders, more than 15, reworked many as well, such as glass and mirrors.

-Added the realtime reflection probes to the scene, also the realtime lighting for day / night.

-Added multiple dynamic components such as TVs, Fire, Water pools, shower, running water, etc…

-Dramatically optimized the performance overall: Runtime more than 95% (FPS), load time ~70%.

-We ended up partnering with V.I.S. in order to distribute VIRay in Mexico. Project Name: Consulting to Accenture • Client: Accenture • Mexico • React Native • 2 months

-Link-1

-Accenture hired me as a mobile consultant to guide their 10-member tech team to create a React Native app to parse in real time information from a photo taken from a smartphone to a form Client: Legends of Learning • USA • Unity 2D WebGL • Scattered in a 3 year timespan

-Link-1 • Link-2

-I created 12 games that are currently active in their platform

-2 of which were considered as Top Grossing in 2017 and 2018

-I am considered as “Top Developer” within Legends of Learning Project Name: Hyundai Website Maintenance • Mexico • JS, HTML, Photoshop • 6 months

-Link-1

-I implemented 100% of their corporate website upon new arrival to the Mexican market.

-I functioned as a marketing agency myself. By editing the car specifications on demand, directly on the public and productive marketing website where all the cars specifications were published. COMPANY: Accenture Mexico

ROLE:Junior Java and Web developer

PERIOD:December 2012 - May 2013

MAIN PROJECTS:

Project Name: Bancomer Net Cash • Client: BBVA Bancomer • Mexico • Web • 6 months

-Link-1

-Migrating services from Bancomer.com to Bancomer NetCash.

-My role was supposed to be only on the designing phase of the 3 modules.

-Upon project completion, leveraging my comprehensive understanding of the module functionalities as the sole developer, I successfully deployed and integrated these modules into Bancomer's global production environment.

Contact this candidate