Software Developer &
AI Engineer

I am passionate about software development and AI engineering. I help companies to enhance their business with AI.

This website serves as a hub for my presence, showcasing my projects and skills. Feel free to explore, and don't hesitate to reach out using my AI assistant (on the right bottom corner) if you'd like to collaborate or connect.

Certificate

I am a certified Associate Data Analyst by DataCamp .

My certificate is valid from Nov 2024 to Nov 2026.

Contact Information

Email

yuxiangpan2019@gmail.com

yuxiang.pan@mail.mcgill.ca (academic)

Mobile

(+1) 514-993-0881 (Canada)

Location

Montreal, Quebec, Canada

Work and Service

Professional experience and contributions

AI Engineer

Smart Automation

Agency Website

Responsibilities:

  • Meet with prospects and design an optimal solution tailored for prospect's business requirements.
  • Deploy AI solutions, design systems, and set up APIs
  • Be versatile, deliver on time, build trustworthy and reliable solutions
  • Retainer maintenance.

Software Engineer

Robert Bosch - UAES

Responsibilities:

  • Lead cross-functional team in the developing enterprise-level software with extendable AI system using Agile and Scrum methodologies.
  • Construct flexible and scalable database for confidential data, with maintenance protocol and action plan to mitigate risks in future development.
  • Develop the software layer: conduct data analysis using Panda and Numpy framework, implement core logics and algorithms, build machine learning model tailored for company data using PyTorch, Tensorflow, and Scikit Learn, develop software's frontend.

Achievements:

  • Implement process automation, reducing manual workload by 60-90%.
  • Add resource recycling to the pipeline, introducing standards.

Programming Language Researcher

McGill University

Responsibilities:

  • Design logic frameworks to replicate real life models (software, hardware, systems, and more!) using functional programming languages.
  • Mechanize normalization by evaluation to enhance model reasoning, reducing model verification speed from days to seconds.

Achievements:

  • The frameworks are applicable to various industrial settings.
  • Automate up to 100% of verification process in the pipeline.

Natural Language Processing Researcher

McGill University

Responsibilities:

  • Research state-of-the-art (2023) natural language processing models for the classification of Chinese characters.
  • Implement and optimize classification machine learning models without high-order semantic logics.

Achievements:

  • Reach +90% model accuracy.

Software Engineer

Anban Tech

Responsibilities:

  • Debug and improve the intelligent fuzziness testing system.

Achievements:

  • Clean noise from data depending on the noise intensity.

Volunteer

Elderly care-taker at Royal Victoria Hospital

Life Guard

Guardian at World Gym Brossard

Asgardian Knight Commander

Protector of seven galaxies :)

Skills

Core competencies and expertise

Full-Stack Software Development

4 years

AI Engineering

2 years

Data Scientist

3 years

ML Scientist

2 years

Computational Modeling

1.5 years

Research and Projects

Academic research and technical projects

Research

Normalization by Evaluation for Simply Typed Lambda Calculus (2024)

Normalization by Evaluation (NbE) is a powerful normalization concept in type theory and functional programming that aims to establish a consistent normal form of representing typable expression. The technique is useful to construct efficient and extendable verifier which can reason about model properties.

This research presents an overview of NbE acting on the model of Simply Typed Lambda Calculus in proof assistant Coq.

Amazing supervisors: Jason Hu, PhD. , Prof. Brigitte Pientka

A Deeper Exploration of Natural Language Processing in Chinese (2023)

We present a Natural Language Processing (NLP) task of text classification for text in Chinese, with a deeper focus on radical and stroke information for individual Chinese character. We utilize multinomial naive bayes classifiers with different selection strategies. Our work proves the effectiveness of the technique, as it outperforms expected accuracy while completely disregards textual semantic logic.

Amazing teammates: Ava Gilmour, Haiqi Zhou

Projects

  • Multiple ML competition experience that scored within top 25% globally, with best ML models ranked 12/993 and 82/4770.
  • Multiple independent projects on GitHub :
  • ML model that can detects hand gesture and translate into simple commonly used phrase or English alphabet using American Sign Language.
  • Cryptographic cipher and attack design.
  • Mechanizing simple shell in C utilizing different memory management methods.
  • Exploration of software development principles and design patterns.
  • Simple video game design and modding.

Technology

Technical proficiencies and tools

AI Tools

Relevance AI, Make, Agentive, Voiceflow, HuggingFace, Google VertexAI, Kaggle Database

Microsoft

Azure, Excel, Power BI, PowerPoint, Word

Python

Daily usage. To implement Machine Learning Model or to build an AI system

Java

Software design and video game modding

C

For low-level high-performance implementation. Building simple shell

R

Efficiently compute various statistical model, such as rank-based non-parametric statistic and regression models

SQL

Data retrieval and manipulation. Neatly structure and classify data while assuring easy and fast access

OCaml

Functional programming, mostly used for research purpose

Coq/Beluga

Construct mathematical models of real life systems (such as software program or security protocol) and reason about their properties to ensure correctness

Languages

Multilingual communication abilities

English

Fluent

French

Fluent

Chinese Mandarin

Native

German

Half-work proficiency

Japanese

Half-work proficiency

Spanish

Beginner

Russian

Beginner

Education

Academic background and qualifications

McGill University

Montreal, Quebec, Canada

Bachelor of Science

Math & Computer Science