Jonathan Moreno Narváez

Jonathan Moreno Narváez

Data Scientist with more than 12 year of general experience. Lover of AI, ML, Optimization, and Data Visualization.

Curriculum Vitae

Corporative Experience

AlgoDriven

Data Scientist Mar. 23 - Now Dubai, United Arab Emirates.
  • Maintained and optimized the used car pricing algorithm, significantly improving its scalability, performance, and monitoring ability.
    • Reduced the error of the main ML algorithm by 33% using boosting and meta-labeling techniques.
  • Utilized predictive modeling and forecasting techniques to accurately predict and forecast used car prices, driving more informed decision-making.
  • Created a few monitoring tools (based on Retool) for internal monitoring and auditing.
  • To improve performance on the inference side, I have established a cache using Redis and scheduled daily jobs to pre-populate certain responses based on similar requests using AWS ECS (Fargate).
  • Developed an embedding and similarity-based model that increased matching ROC AUC by 30% over a query-based baseline, aligning online car listings with a proprietary database of over 23,000 vehicles.
  • Set up the foundation of MLOps to experiment and operationalize new ML models faster.
    • Introduced software engineering practices such as Unit Test in the CI/CD pipeline in order to ensure the quality of the ML models before deployment.
    • Developed and maintained the CI/CD pipeline for the current pricing AI (using Bitbucket, AWS ECR, AWS EC2, and Docker Swarm) which enhanced the speed of production releases and minimized downtime.
  • Gave support to Ad Hoc analysis related to Residual Value and insights on the UAE used car market.

AHOY

Data Scientist Mar. 22 – Mar. 23 Dubai, United Arab Emirates.
  • Helped the Data Analytics/Operation Research team to grow from 1 to 8 members.
  • Contributed to the API design and development of the Multi Stop API for AHOY AMS, solving an NP-hard problem: Vehicle Routing Problem using ML capabilities in different layers. We optimize transportation costs, distance, and time while ensuring customer deliveries are completed on time and several other constraints. The solution has helped AHOY to position itself as a reliable routing solution for Last Mile Delivery services in MENA.
  • Developed and maintained the CI/CD pipeline of the Routing Engine using FastAPI, Azure Functions, Docker Containers, and Azure DevOps.

Freelancer – Self-Employed

Data Scientist Oct. 21 – Mar. 22 Dubai, United Arab Emirates.
  • Funka: This pro-bono project used SAS Viya to develop an object detection system. We were recognized by SAS with three awards in the 2022 Hackathon. The underlying model was YOLO V3. I was responsible for designing the model’s architecture, the model’s training pipeline and ensuring that it was properly deployed and integrated.
  • Negotiations.com: Data Analysis and general guidance on introducing AI/ML capabilities to the negotiation’s platform.
  • IncreasIA: Developing Deep Learning models to assist the trading operation in FX markets. I also had to ensure that the models were robust and reliable and that the data was properly organized and pre-processed to maximize the accuracy of the models.
  • Sonria.com.co: Developed a predictive model that can accurately identify customers who are likely to churn.

iBLOXX Capital AG

Financial Data Scientist Jun. 21 – Sept. 21 Dubai, United Arab Emirates.
  • Data collected and properly stored from data providers to be effectively used in strategy development.
  • Crypto-based trading strategies developed in Zorro Project and Python.
  • Cython application in proprietary backtesting engine managing tick level data improving up to 500x the speed of analysis.

Globant

Data Scientist Semi Senior Aug. 19 – May. 21 Bogotá D.C., Colombia.
  • Data Visualization (Reporting and Dashboards) in D3.js and React.js (Some examples)
  • PoC on Computer Vision (OpenCV and Haar Cascade for Facial Detection)
  • For a Consulting client: PoC on Pricing & Forecasting. To solve this problem, we used all the Data Modelling Pipeline to end up in the Prescriptive phase. Some topics handle for this project are:
  • Entity Relationship Diagram and SQL Server.
  • Exploratory Data Analysis and Feature Engineering.
  • Natural Language Processing.
  • Unsupervised models – Clustering and Supervised models - Regression.
  • Recommendation Engine (Memory-based and Model-based algorithms)
  • Model deployment with Flask via API.
  • Mathematical optimization solver with Python.

Presidencia de la República de Colombia

Lead Data Scientist Sept. 18 – Jun. 19 Bogotá D.C., Colombia.
  • Working side-by-side with stakeholders in data-driven Government initiatives.
  • Descriptive statistics on Social Media data in Official accounts.
  • Data mining, data processing and data visualization support for decision-making over communications’ strategies.
  • Social Media Listening - Facebook API and Twitter API connection.
  • Network analysis (Gephi, Networkx - Python, and R).
  • Poll analysis.
  • Promoting and managing the implementation of Big Data, Machine Learning and AI solutions for projects at the Proof-Of-Concept stage within the Government.
  • Sentiment Analysis in-house development with python (NLTK, Textblob, Scikit-Learn).
  • Lead the development of new mechanisms for Civic Engagement where data and new technologies generate trust and bring the Government with its citizens closer.
  • Data Visualization: Exploring infogram and tableau. Implemented solutions on D3.js, plotly, matplotlib.
  • Dashboards for data visualization on PowerBI.
  • Use data analysis tools & SQL (postgres) to analyze and manipulate large data sets.
  • Databases Management on Postgres/MySQL (localhost and AWS RDS).
  • Web Apps Development. Django, Flask, and Ruby on Rails. (www.pactoporcolombia.gov.co and www.energiaevoluciona.org)
  • E-Government: Web apps (Proof-Of-Concept) testing Civic Engagement, Open Government, and Data Analytics hosted on AWS (Linux Amazon, Ubuntu, MacOS).

Colombian Presidential Campaign – DUQUE 2018-2022

Data Scientist Apr. 18 – Aug. 18 Bogotá D.C., Colombia.
  • Digital Marketing and decision-making support: Data Analysis for a Colombian Political Presidential Campaign applying Python, R, Web Crawling and Scraping, MySQL, Anaconda, Spyder, RStudio, APIs for Social Networks, VBA-Excel, Google Data Studio for visualization, Google Cloud and Linux.
  • Pareto Analysis and Micro-Targeting models. Polls’ analysis.
  • E-Government strategy evaluation. @EGovernmentCOL1
  • Social Media Listening.

Wood

Process Engineer Aug. 17 – to Mar. 18 Bogotá D.C., Colombia
  • P&ID checker for Process, Plant, and Industry projects. Managed, coordinated and executed projects of budgets ~10,000 USD per month at Work-sharing unit with US offices: Monomer Expansion Plant, Chemical Plants, Gas Station Plants and Terminals, Valve Maps development for Gas Complex. Hydraulic analysis. Basic and Detail Engineering: P&ID design and Process deliverables execution for Colombian Polymer plant.
Process Engineer Jan. 15 – Apr. 16 Bogotá D.C., Colombia
  • Deliverable execution of process equipment datasheet, mass and energy balances, P&ID and PFD.
  • Line sizing, valve estimation, and major facility equipment design. Hydraulic analysis. Relief devices calculations. PSV studies. Pump systems assessment.
Process Safety Engineer July 13 – Jan. 15 Bogotá D.C., Colombia
  • Decision making via modelling and simulation (DNV PHAST, VBA and Stoner Pipeline Simulator).
  • Risk-based evaluations such as High Consequence Areas and Quantitative Risk Assessments. Support for developing Reliability, Availability and Maintainability Analysis.

Enterpreneurship Experience

Data Enigma

From Jan. 20 - Now Bogotá D.C., Colombia.

Follow me on Twitter: @EnigmaData

  • Social Network Analysis for Twitter, Facebook, and Instagram. Applied Natural Language Processing techniques along with Machine Learning to Sentiment Analysis projects.

  • Premex - Asimetrix

    Data Scientist - Freelance Jul. 20 – Jan. 21 Bogotá D.C., Colombia
    • Data Visualization (Plotly)
    • Time series analysis and forecasting.
    • Machine Learning based product development.
    • Dash development.

  • Scalable Path

    Data Scientist (Remote) May. 18 – Apr. 19 Bogotá D.C., Colombia.
    • Involved in projects with Rehinged:
      • Monte Carlo Analysis and usage of Data Science toolkits, such as R, NumPy, Python, Dash, etc
      • Data visualization tools, such as D3.js, ggplot, plot.ly, matplotlib etc.
      • Nonconventional data viz structures: Networkx, Sankey.
      • Data crawling for market analysis and discovering new investment opportunities from unstructured data sources.
      • Databases design and. Connection with data lakes and data warehouses.
      • Google API connection for getting information around given locations.

  • Imperial Finance

    Quantitative Researcher (Remote Work) Nov. 17 – to Feb. 18 Bogotá D.C., Colombia.
    • Implement trading models involving Time Series classification and prediction using Machine Learning and Deep Learning (CNN, LSTM) algorithms achieving nearly 55% Out-of-Sample. Generate new ideas for trading models.
    • Code and test models’ performance. Applied enhancement techniques for python codes (cython) and memory handling for 30 GB of financial data. Reduced crunching time consume by 15x with cython implementation.
    • Execute projects on Google Cloud Platform including BigQuery, Google Cloud Storage, and Google Datalab.
    • Major FX pairs microstructure analysis and anomalies detection on Level II-DOM Market data.

  • DecisionWare LTDA

    Process Optimization Engineer (Remote work) June 16 – to Feb. 17 Bogotá D.C., Colombia.
    • Managed and executed projects with budget ~20,000 USD: Mathematical optimization model for improving Energy Efficiency through Optimal Blending, Planning, and Scheduling. (GAMS application) Nonlinear programming; Supply chain optimization through mathematical programming. (GAMS application) Linear programming.

  • Wood

    Consultant (Remote work) - Oleoducto de los Llanos Orientales (ODL) May 16 – to Dec. 16 Bogotá D.C., Colombia
    • Algorithm making and development of optimization model for 250 km pipeline operation.
    • Programming on VBA to perform HCA (High Consequence Areas) Risk Analysis and Assessment of pipelines based on ASME-B31G.

Trading Enigma

From Mar. 17 - Sept. 21 Bogotá D.C., Colombia.

Follow me on Twitter: @TradingEnigma

  • Trading Strategy Development for different markets on R, Python, Metatrader 5 (MQL5), Multicharts and TradeStation (EasyLanguage). BuildAlpha User (Data Mining).
  • Financial modelling: Time series analysis (ARIMA, GARCH, etc). Black-Scholes applied to options, Monte Carlo, Markov Chain, some techniques of Machine Learning.
  • Executed a research project on “Pair Trading: Statistical Arbitrage on Cash Stocks” in which were applied tools like: statistical analysis, Object Oriented Back testing (QSTrader) and Monte Carlo Analysis.
  • Worked with full cycle of Algo trading development which includes:
    • Data Management and Technical Analysis.
    • Algorithmic Development.
    • Programming.
    • Backtesting and Optimization.
    • Trade execution and Risk Management.
    • Portfolio Analysis.
    • Live Execution Reporting.
    • Position Management.
    • Exposure Monitor to different Risk Factors.
  • Option active trading managing own money.

Education

  • Master of Science in Chemical Engineering from Jan. 12 – to June 13
      Universidad de los Andes, Bogotá D.C., Colombia.
  • Bachelor of Engineering, Chemical Engineering with Minor in Applied Mathematics from Jan. 08 – to Dec. 11
      Universidad de los Andes, Bogotá D.C., Colombia.

Teaching Experience

  • Universidad Jorge Tadeo Lozano

    Adjunct Professor July 15 – Dec. 15 Bogotá D.C., Colombia.

  • Lecturer on Chemical Processes Optimization. Topics included: Matlab/Octave/Scilab/GAMS programming basics, Numerical Methods in Chemical Engineering; Structure and solution methods for Optimization problems (Unconstrained, LP, NLP, MINLP, etc.)

  • Universidad de la Sabana

    Adjunct Professor July 15 – Dec. 15 Bogotá D.C., Colombia.

  • Lecturer in Modelling and Simulation in Chemical Engineering. Topics included: Matlab programming basics, general mathematical modelling and its applications on thermodynamics, transport phenomena and control of dynamic systems.

  • Universidad de los Andes

    Teaching Assistant Jan. 12 – June 13 Bogotá D.C., Colombia.

  • Lecturer in Modelling and Simulation in Chemical Engineering. Topics included: Numerical Methods, Matlab programming basics, mathematical programming, thermodynamics, transport phenomena and unit operations.

Computer Skills

Languages Known Databases Tools Frameworks Version Control System Cloud Knowledge
Python, R, GAMS, Excel-VBA, Xpress, VBA, Ruby (Entry Level), Javascript (Entry Level) MS SQL Server, MySql, PostgreSQL LaTex, Microsoft Office, Matlab, Metatrader 4/5 & MQL4/5, Interactive Brokers API, OANDA API, BuildAlpha, Multicharts, TradeStation Tensorflow, Pytorch, React.js, D3.js, Plotly - Dash, RapidAPI, BeautifulSoup, Scikit-Learn, Matplotlib, Selenium, OpenCV, NLTK, Transforrmers, Flask, Pandas, Numpy, Dask, Tweepy, Keras, XGBoost Git AWS ElasticBeanstalk, AWS EC2, AWS RDS, AWS Lambda, Azure Functions, Google BigQuery, Google Data Studio, PowerBI, Google Cloud Datalab, Google AI Platform.

Activities and Interests

LANGUAGES: Spanish (Native Language), English (Advanced), German (Entry Level), French (Entry Level).

ACTIVITIES: Programming, football, hiking.

INTERESTS: Systematic Trading, Investment Management, Machine Learning, Artificial Intelligence, Applied Mathematics on Finance, Modeling, Mathematical Programming, Operation Research, Data Analysis, Algorithmic Trading, Quantitative Trading Analysis, Stochastic Processes, IoT, Big Data, Web Apps, Data Visualization.

Updated: