Quality control engineer

While working in the field of banknote quality control, I focused on applying machine learning techniques to support watermark approval, using the Anomalib library for both supervised and unsupervised anomaly detection. Alongside this, I designed a GUI application that connects with the Datacolor45IR spectrophotometer to perform CIE Lab and CMYK density calculations, tailored for deployment in certified European printing facilities. I also took part in on-site inspections of printed banknotes and helped draft technical documentation aimed at both internal teams and external partners.

Data scientist

During my time at Kaavio, I worked with technologies like CrewAI and Langchain to build an LLM-powered search engine. This engine used a crew of agents to retrieve, validate, and summarize company information through keyword-based queries. The work done inspired me to develop a series of personal projects that helped me practice and gain a better understanding of these technologies. Personal projects that I developed involved the use of LangChain to build RAG agents or agents capable to communicate with SQL databases.

Process engineer

During my time at Cooper Standard's mixing facility, I held the position of a ML Engineer, where my main focus was on building neural networks to forecast experimental parameters derived from the formula's ingredients. This initiative was aimed at reducing waste and optimizing our production workflow. The models I designed consistently demonstrated an impressive accuracy rate of up to 90%, leading to a significant 2% decrease in waste and considerable cost savings in production.

Skills enhanced during the job:

• SQL: Utilized SQL Server Management Studio for data retrieval from a 20 GB database. Leveraged advanced SQL functionalities including CTE, Pivots, and window functions
• Python: Employed PyTorch framework for developing artificial neural networks (ANNs). Incorporated advanced loss functions and weight decay regularization technique for model optimization and generalization.

Fulltime student

As a graduate student, I took on an active role as a Teaching Assistant while simultaneously delving into my research project. Some of the courses where I served as a TA included:

• Chemical Process Modelling and Simulation - CHE 520

  Winter 2024

  As the primary and only teaching assistant, my duties included grading assignments, hosting office hours, and leading tutorials to teach students advanced features of ASPEN-Plus. Moreover, I offered guidance on programming shortcut methods for unit operations using Python, covering topics such as OOP and root finding.

• Computer literacy and programming for chemical engineers - CHE 120

  Fall 2023

  My responsibilities as a teaching assistant included grading assignments, conducting tutorials twice a week to provide hands-on programming instruction to students, and holding office hours twice a week to address student inquiries.

• Design Studio 2 - CHE 181

  Spring 2023

  My responsabilities as a teaching assistant were grading quizzes and final reports. Final reports involved students' discussion of a practical experiment using Arduino to measure melting points of water and saltwater

Research assistant

In my final semester of undergraduate studies, I dived deep into the integration of Python with the chemical simulation suite ASPEN-Plus, aiming to develop a framework for real-time and model-free optimization. Notably, I applied this framework to optimize both the operating pressure and the composition of the working fluid within an Organic Rankine Cycle (ORC), using advanced genetic algorithms.

Some of the libraries used in this project were:

• Pymoo: A Python library featuring pre-built genetic algorithms, prominently NSGA-II.
• NumPy: A fundamental package for numerical computations in Python, facilitating array operations and mathematical functions.
• Win32 COM: A technology used to establish connections with applications, such as ASPEN, for process interaction and automation.

• Go back to About me