Kasra Darvishi

Kasra Darvishi Email and Phone Number

Data Scientist @ HEC Montréal
St. Catharines, ON, CA
Kasra Darvishi's Location
St. Catharines, Ontario, Canada, Canada
About Kasra Darvishi

Data Scientist with over three years of research and industry experience, focusing on Natural Language Processing (NLP), Anomaly Detection, and Transfer Learning.• Developed and deployed machine learning solutions across the full lifecycle, from data ingestion and transformation to inference, using both classical and deep learning methods.• Presented complex analytical results to various stakeholders, helping to enhance strategic decisions in both industrial and academic environments.• Skilled in data analysis, cleaning, and visualization using tools like Pandas, Scikit-learn, Jupyter Notebooks, Numpy, and Scipy.• Earned an MSc in Computer Engineering, specializing in ML, NLP, and Information Retrieval from Amirkabir University of Technology. • Published research in prestigious forums like the International Conference on Software Engineering (ICSE).Technical skills:- Deep Learning: PyTorch, TensorFlow, Keras- Language Models: Extensive experience in fine-tuning and prompt engineering- Hardware Acceleration: Proficient in using CUDA for performance optimization- Version Control: git CLI

Kasra Darvishi's Current Company Details
HEC Montréal

Hec Montréal

View
Data Scientist
St. Catharines, ON, CA
Kasra Darvishi Work Experience Details
  • Hec Montréal
    Data Scientist
    Hec Montréal
    St. Catharines, On, Ca
  • Lingup
    Machine Learning Engineer
    Lingup Jul 2024 - Present
    Ontario, Canada
  • Goodman School Of Business At Brock University
    Data Scientist
    Goodman School Of Business At Brock University Jun 2024 - Present
    Ontario, Canada
  • Niagara Health
    Monarch Project Support
    Niagara Health Oct 2024 - Jan 2025
    St. Catharines, Ontario, Canada
  • Brock University
    Machine Learning Researcher
    Brock University Jan 2023 - Mar 2024
    Ontario, Canada
    • Developed a novel language model-based method for Linux kernel tracing, funded by Ciena, minimizing storage and processing overhead by adapting tracer settings based on real-time system behavior analysis. • Optimized an ML pipeline that extracts, cleans, and preprocesses trace data, achieving 800 requests/second throughput on a multi-GPU setup.• Validated the method by collecting and analyzing data from a web server processing over 2 million requests. • Achieved a 70% reduction in tracing overhead while maintaining 96% accuracy by targeted data collection. • Employed a customized vectorization technique to classify abnormal web requests into 7 categories with 99% accuracy, streamlining automated anomaly detection and root-cause analysis. • Authored a paper on our system monitoring approach, accepted at the prestigious ICSE.• Implemented advanced data visualizations for another comprehensive paper currently under review, enhancing data interpretation and supporting analytical insights.
  • Fanap
    Machine Learning Engineer
    Fanap Jun 2022 - Jan 2023
    Tehran, Tehran Province, Iran
    • Played a key role in the development of a chatbot system prototype.• Enhanced and optimized intent detection and named-entity recognition (NER) modules by fine-tuning transformers like BERT, improving functionality by up to 15% and cutting response latency by 50%.• Implemented sentiment analysis algorithms for the demonstration of customer conversation analysis, effectively identifying satisfaction levels and pinpointing unresolved issues with 90% accuracy. • Designed and integrated a scalable information retrieval system and question-answering mechanism into the prototype. • Curated specialized datasets through web scraping from 50+ sources for initial model training and evaluation.
  • Amirkabir University Of Technology - Tehran Polytechnic
    Graduate Research Assistant
    Amirkabir University Of Technology - Tehran Polytechnic Nov 2020 - Dec 2022
    Tehran, Tehran Province, Iran
    • Managed quality assurance process and conducted detailed data analysis for the first crowdsourced Reading Comprehension dataset in Persian with 80,000 questions, ensuring data integrity and uncovering key insights. • Documented the study’s methodology, leading to a publication in the Computer Speech & Language journal . • Introduced an innovative multi-task model to enhance question answering systems for low-resource languages, specifically increasing the detection accuracy of unanswerable questions from 92% to 94%.
  • Amirkabir University Of Technology - Tehran Polytechnic
    Graduate Teaching Assistant
    Amirkabir University Of Technology - Tehran Polytechnic Feb 2022 - Jun 2022
    Tehran, Tehran Province, Iran
    Courses:- Information Retrieval (Graduate) - Winter 2022
  • Amirkabir University Of Technology - Tehran Polytechnic
    Teaching Assistant
    Amirkabir University Of Technology - Tehran Polytechnic Sep 2018 - Jun 2019
    Tehran, Tehran Province, Iran
    Courses:- Multicore Programming - Winter 2019- Computer-Aided Digital System Design - Winter 2019- Digital Electronics - Fall 2018
  • Institute For Research In Fundamental Sciences (Ipm)
    Research Intern
    Institute For Research In Fundamental Sciences (Ipm) Jun 2018 - Sep 2018
    Tehran, Tehran Province, Iran

Kasra Darvishi Education Details

Frequently Asked Questions about Kasra Darvishi

What company does Kasra Darvishi work for?

Kasra Darvishi works for Hec Montréal

What is Kasra Darvishi's role at the current company?

Kasra Darvishi's current role is Data Scientist.

What schools did Kasra Darvishi attend?

Kasra Darvishi attended Amirkabir University Of Technology - Tehran Polytechnic, Amirkabir University Of Technology - Tehran Polytechnic.

Not the Kasra Darvishi you were looking for?

Free Chrome Extension

Find emails, phones & company data instantly

Find verified emails from LinkedIn profiles
Get direct phone numbers & mobile contacts
Access company data & employee information
Works directly on LinkedIn - no copy/paste needed
Get Chrome Extension - Free

Aero Online

Your AI prospecting assistant

Download 750 million emails and 100 million phone numbers

Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.