Logo
Lucian – Python, SQL, MySQL, experts in Lemon.io

Lucian

From Romania (GMT+3)

flag
Data Scientist
Hire developer
7 years of commercial experience
Customer support
E-commerce
Fintech
Logistics
Retail
AI software
Lemon.io stats
2
projects done
370
hours worked
Open
to new offers

Lucian – Python, SQL, MySQL

Lucian is a passionate Data Scientist with over 7 years in the industry, primarily focusing on Python, SQL, MySQL. He constantly working on the improvement and enhancement of his skillset by implementing various educational resources and methods into his daily routine.

Main technologies
Python
4 years
SQL
2.5 years
MySQL
2.5 years
Additional skills
AWS
BigQuery
XML
Apache Spark
Scala
C++
Bash
Microsoft SQL Server
GCP
NumPy
Pandas
Scikit-learn
Ready to start
ASAP
Direct hire
Potentially possible

Experience Highlights

Team Lead Quantitative Researcher
Apr 2022 - Apr 20231 year
Project Overview

Their process involves distinctive, proprietary machine learning technology that aims to natively incorporate transaction costs, diversification, and risk constraints to produce persistent, high quality returns.

Skeleton
Skeleton
Skeleton
Responsibilities:
  • Managed a team of 3 quants, conducted interviews, set research directions, planning & milestones
  • Designed systematic daily frequency long-short equity strategies on a broad range of data categories and various modules for cost optimization, turnover reduction, liquidity enhancement; performed alpha research on a set of topics including factors, insiders, shorts, fundamentals, sentiment, and price-volume
  • Developed a complex framework for signal generation, including features for data ingestion, robustness checks, a wide range of transformations, an operators & parameters optimizer, visualization tools, and other complementary instruments used in alpha research
  • Created grouping datasets using ML algorithms for hedging risk at a granular level within the universe
Project Tech stack:
Python
SQL
GCP
Machine learning
Algorithms and Data Structures
Data Science
Data Modeling
Data Scientist
Aug 2021 - Oct 20212 months
Project Overview

The purpose of this project was to create an application that delivers estimates for the cumulative sums of ordered quantities for about 90 locations that contain one or multiple locker delivery points. These estimates serve as the input for the responsible business teams, which transform them using estimated conversion rates into actual orders per locker and later make decisions of increasing capacities or blocking certain delivery locations based on how the numbers evolve in real-time. The project was deployed using Flask-based endpoints, which were triggered by the business teams every 15 minutes, while the prediction algorithm was the Passive Aggressive Regressor from scikit-learn.

Skeleton
Skeleton
Skeleton
Responsibilities:
  • Was the main point of contact for any question or analysis request coming from partner departments
  • Derived the machine learning project strategy and timeline, proposing an initial set of features, running different model development iterations
  • Came up with a final model proposition
  • Performed integrated testing along with other departments
  • Contributed with different reports and ad-hoc analyses for the business stakeholders
Project Tech stack:
Python
MySQL
Microsoft SQL Server
Bash
Flask
Data Scientist
Jun 2021 - Sep 20213 months
Project Overview

Machine learning solution for the fintech solution of the Finance Department aiming to offer credit-related products to customers in order to boost sales and diversify the payment channels and schedules. This piece is represented by a credit scoring model that considers a large set of client-related features (based on the customer relationship history with the firm: order characteristics, client behavior, demographics etc.) and outputs a score (similar to a probability of default, the likeliness of not repaying the credit amount, taking advantage of a Boosting Classifier algorithm) corresponding to each client, enabling the business to derive client cohorts and offer custom solutions tailored by common characteristics and overall risk appetite.

The overall project translated to an additional 6-figures EUR revenue per month coming via this financial product (more than 100k individual clients choosing this payment method over a 3-months period), representing a success for my team and for the company strategy.

Skeleton
Skeleton
Skeleton
Responsibilities:
  • Was the main point of contact for any question or analysis request coming from partner departments
  • Contributed to the initial business case with different pieces of analysis
  • Derived machine learning project strategy and timeline
  • Proposed an initial set of features
  • Constructed the data pipelines in order to build the main database
  • Ran different model development iterations
  • Came up with a final model proposition, deriving the retrain and inference flows
  • Suggested the model output interpretation
  • Contributed with different reports and ad-hoc analyses for the business stakeholders
Project Tech stack:
Python
Microsoft SQL Server
MySQL
Apache Spark
Scala
Dremio
Superset.
Quantitative Researcher
Jan 2018 - Aug 20213 years 6 months
Project Overview

As a quantitative researcher, Lucian developed tens of algorithmic daily re-balancing long-short neutral trading signals for multiple markets (US, EUR, Asia), on different asset classes (incl. Equities, ETFs), using a wide range of data categories (e.g. fundamental, analyst, financial derivatives, sentiment, news, etc.).

Such alphas were designed to make time-series predictions in order to take advantage of the most recent changes in the data and automatically adjust the vector positions of the trading strategies so that within a very short time they take advantage of market imperfections and generate profits. The trading signals were aimed to be used by the portfolio managers within the hedge fund (that were filtering by out-of-sample performance metrics like Sharpe Ratio, rate of return, turnover, liquidity, maximum drawdown, correlation, etc.) which were integrating such signals within their flag strategies.

Skeleton
Skeleton
Skeleton
Responsibilities:
  • Did literature reviews
  • Suggested datasets or necessary fields in order to test the main hypothesis
  • Designed the algorithms
  • Ran back-testing
  • Developed and performed several checks in order to prevent overfitting
  • Built alpha operations for twisting the alpha signals in order to improve their performance, liquidity, turnover and submitted the alpha signals to portfolio managers.
Project Tech stack:
Python
Google BigQuery
C++
XML
Bash
Main Data Scientist
Oct 2020 - Jan 20213 months
Project Overview

The objective of the solution was to monitor the 11 seller indicators via which they can operate on the marketplace (rate of return, parcel tracking, customer reviews, etc.) and have the ability to forecast between 1 and 3 months whether the indicator's value would drop below a certain threshold and reach the yellow/red flag, the stage where the seller might be suspended or eliminated from the marketplace, so that either the firm or the individual seller can take action and prevent such scenarios, reducing the overall turnover rate. As the issue was mainly a time-series forecasting task, the proposed solution represented a combination of a boosting regressor and a DeepAR algorithm (a combination of a recurrent neural network and an ARIMA classical forecasting method). The model helped in dropping the elimination rate by ~7% in the first 9 months of operating.

Skeleton
Skeleton
Skeleton
Responsibilities:

Lucian was responsible for every request in the machine learning-related area. His responsibilities included designing a preliminary algorithm solution and timeline, suggesting an initial set of features, building the data engineering pipelines, assembling the main training data, performing the model development, deriving the model retrain and inference flows as well as answering any business requests in terms of ad-hoc analyses, reports.

Project Tech stack:
Python
Microsoft SQL Server
MySQL
Apache Spark
Scala
Data Science
DevOps
SQL
Machine leaning
Main Data Scientist
May 2020 - Aug 20203 months
Project Overview

Working as a Data Scientist for one of the largest e-commerce companies in South-Eastern Europe, Lucian was the main developer of a machine learning solution for the Customer Care department in order to enable them to filter every interaction with the customers by the likelihood of having another contact with the client on the same matter within a 7-days period of time, helping them increasing the first-contact-resolution KPI from 59% to 72% within a 3-months period, reducing the load of the Customer Care department and reducing their costs. The proposed machine learning solution, a binary classification method, was taking advantage of a large set of features about the client, the corresponding order, and the subject matter, generating a probability of a new interaction within the specified time interval, enabling the agents to actively approach clients and take action so that the customers are aware of the steps taken by the company in solving their matters without needing to inquire again about the issue.

Skeleton
Skeleton
Skeleton
Responsibilities:
  • Was responsible for every request in the machine learning-related area
  • Designed a preliminary algorithm solution and timeline
  • Suggested an initial set of features
  • Built the data engineering pipelines
  • Assembled the main training data
  • Performed the model development
  • Derived the model retrain and inference flows
  • Answered any business requests in terms of ad-hoc analyses or reports
Project Tech stack:
Python
Microsoft SQL Server
MySQL
Scala
Bash
Data Modeling
Data Science
Machine leaning

Education

2016
Mathematics, Operational Research, Statistics & Economics (Actuarial & Financial Mathematics Stream)
Master's

Copyright © 2024 lemon.io. All rights reserved.