Teja is available for hire

Teja Arikati

Verified Expert in Engineering

Data Scientist and Developer

Location

Philadelphia, United States

Toptal Member Since

February 28, 2023

Teja is a versatile data scientist with over five years of experience and a postgraduate operations research (OR) degree from Cornell. His experience ranges from implementing end-to-end machine learning solutions to leading a team of data scientists and analysts. Teja is always ready to deliver production-quality results to any data science project.

Data Visualization Data Analysis Data Analytics Exploratory Data Analysis Machine Learning Natural Language Processing (NLP)Time Series Analysis Seaborn PySpark SQL Scikit-learn Python Linear Programming Artificial Intelligence (AI)Visual Studio Code (VS Code)StatsModels Gurobi

Portfolio

Self-Employed

Python, ARIMA, LSTM, Car Rental, Pricing Models, Demand Planning, SARIMA...

Citi

Generative Pre-trained Transformers (GPT), GPT...

Experience

Visual Studio Code (VS Code) - 5 years SQL - 5 years Scikit-learn - 5 years PySpark - 5 years Machine Learning - 5 years Python - 5 years GPT - 5 years Natural Language Processing (NLP) - 5 years

Availability

Part-time

Preferred Environment

Jupyter Notebook, Visual Studio Code (VS Code), PySpark, SQL, Scikit-learn, Apache Hive, Gurobi

The most amazing...

...thing I've achieved is developing and productionizing a company's first machine learning solution, which helped decrease manual labor costs by 25%.

Work Experience

Car Rental Fleet Owner

2021 - PRESENT

Self-Employed

Built forecasting pricing models using ARIMA and LSTMs to forecast demand and optimal pricing of rental cars, increasing monthly revenue by 15%.
Determined the best cars to include in the fleet by calculating the vehicles' time value and expected rental income by car.
Built a passive income stream by hiring employees to manage a self-funded fleet of seven Jeeps on Maui.

Technologies: Python, ARIMA, LSTM, Car Rental, Pricing Models, Demand Planning, SARIMA, TensorFlow, Data Visualization, Data Aggregation, Data Analysis, Data Analytics, Exploratory Data Analysis, Seaborn

Senior Data Scientist

2018 - 2022

Citi

Led the research and Python production deployment of a boosted-tree classifier to calculate the risk score of an email or chat, reducing false positives by 50% and yearly manual review costs by $500 thousand.
Developed and deployed NLP topic detection models, which improved recall by three times while reducing time to deployment by 70% through implementing few-shot learning, active learning, and robust labeling techniques.
Developed an 80% accurate NLP sentence similarity engine by utilizing a domain-specific Word2vec model trained on 300GB of e-communications with PySpark ML libraries.
Reduced compute and storage requirements by 50% in the Hive big data environment by building and optimizing terabyte-sized data stores using PySpark.
Developed and deployed real-time model monitoring tools, which reduced model deviance detection time from two days to 15 minutes.
Mentored and led a team of data analysts and data scientists through the development of standardized data science practices and documentation frameworks.
Automated the business rules extraction from unstructured text utilizing data-driven rule extraction methods such as RuleFit and decision trees.
Researched and fine-tuned BERT to identify specific low-occurrence topics in text and showed improved recall by 30%.

Technologies: Natural Language Processing (NLP), GPT, Generative Pre-trained Transformers (GPT), Python, Apache Hive, PySpark, Jupyter Notebook, Visual Studio Code (VS Code), Machine Learning, Active Learning, Artificial Intelligence (AI), Forecasting, Deep Learning, Rulefit, Decision Trees, Business Rules Engine, BERT, TensorFlow, Decision Modeling, Data-driven Decision-making, Data Visualization, Data Aggregation, Data Analysis, Data Analytics, Exploratory Data Analysis, Seaborn, SQL, Data Science

Experience

Image Search Optimization

This was the final machine learning project for Cornell Tech. I developed an image-search model to predict a picture tag. It extracted image embeddings from ImageNet and Word2vec embeddings from search phrases. I achieved top a 10% score in the class.

Manufacturing Flow Shop Optimization Program

A Python program used by a local CNC machine shop to optimize the scheduling of machine jobs. I worked with the local shop to identify areas of improvement and identified long downtimes of machines on the shop floor. I developed a scheduling system for the shop utilizing Python and linear programming in Gurobi, decreasing downtimes by 20% and increasing throughput by 15%.

Hummus Company Delivery Optimization and Demand Forecasting

I worked with Baba Hummus to improve company operations and identify significant supply chain frustrations.

I built a Python program to calculate optimal daily delivery routes to customers. Using historical delivery data, I modeled customer reliability using logistic regression with 95% accuracy. Utilizing customer reliability scores for different time windows and the OpenStreetMap API for navigation, I modeled a delivery optimization problem using Python and Gurobi. Using the optimal daily delivery schedule output from this program, Baba Hummus completed 20% more deliveries and reduced spoilage costs by 10%.

In addition to this optimization project, I developed a demand forecasting model to predict future demand for 10 different hummus products. I built ARIMA and SARIMA models in Python to predict daily and weekly demand of hummus. In testing, MAPE improved from 10% to 2%.

Skills

Languages

SQL, Python

Frameworks

Business Rules Engine

Libraries/APIs

PySpark, Scikit-learn, XGBoost, Matplotlib, Rulefit, LSTM, TensorFlow

Tools

StatsModels, Seaborn, Gurobi

Paradigms

Data Science, Linear Programming

Platforms

Jupyter Notebook, Visual Studio Code (VS Code)

Other

Machine Learning, Supply Chain Optimization, Natural Language Processing (NLP), ARIMA, Forecasting, ARIMA Models, Decision Trees, Linear Optimization, Logistic Regression, Time Series Analysis, Data-driven Decision-making, Data Visualization, Data Aggregation, Data Analysis, Data Analytics, Exploratory Data Analysis, GPT, Generative Pre-trained Transformers (GPT), Operations Research, Optimization, Process Improvement, Active Learning, Artificial Intelligence (AI), Supply Chain Management (SCM), SARIMA, Pricing Models, Demand Planning, BERT, Mixed-integer Linear Programming, Logistics, Decision Modeling, Deep Learning, Car Rental

Storage

Apache Hive

Education

2017 - 2018

Master's Degree in Informatics and Applied Mathematics

Cornell Tech - New York, NY, United States

2013 - 2017

Bachelor's Degree in Industrial Engineering

California Polytechnic University - San Luis Obispo, California, United States

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring