Curriculum Vitae

Basics

Name Omid Vaheb
Label ML Engineer
Email ovaheb@gmail.com

Work

  • 2025.01 - Present
    ML Engineer @ REDspace
    • Reduced Automatic Speech Recognition (ASR) word error rate & hallucination up to 70%, leveraging contextual data and GPT-4o for generative error correction (EC) in Mockingbird product.
    • Deployed multimodal EC solution on top of OpenAI Whisper, achieving SOTA performance as a PoC.
  • 2024.09 - Present
    ML Associate @ Vector Institute
    • Developed and deployed deep learning solutions for SMEs, collaborating with cross-functional teams to integrate AI into real-world applications while ensuring scalability and efficiency.
    • Gained hands-on expertise in RAG, LLM fine-tuning & alignment for enterprise AI optimization.
  • 2023.06 - 2023.09
    Applied Scientist @ National Research Council Canada
    • Reduced MAE by 15% in astronomical image reconstruction by optimizing Restormer Transformer and U-Net architecture. Implemented in PyTorch and deployed on a multi-GPU cluster.
    • Generated large-scale synthetic data from real galaxy morphologies, reducing CPU synthesis time by 20%. Developed preprocessing pipelines for raw telescope data.
  • 2022.09 - 2024.12
    Research Assistant @ University of Toronto
    • Improved astronomical image denoising by 15dB PSNR, boosting object detection by 7%, and cutting observation costs by 66%.
    • Optimized data loaders, reducing memory usage by 80%, execution time by 60%, and limiting disk access, accelerating model training.
    • Developed an unsupervised denoising framework combining N2N and SURE loss, translating findings to non-technical stakeholders.
  • 2021.04 - 2022.08
    Research Assistant @ Computational Audio-Vision Lab
    • Achieved SOTA 92% accuracy in autism detection from infants’ crying voice by fine-tuning Google YAMNet and using an ensemble classifier in TensorFlow.
    • Attained 99.5% similarity & consistency with human annotations by fine-tuning YAMNet and using energy-thresholding for cleaning voice recordings.
  • 2020.05 - 2021.04
    Data Scientist @ Virasad Co.
    • Reduced maintenance costs by 10% with real-time anomaly detection using LSTM and ARIMA for production line time series forecasting.

Education

  • 2022.09 - 2024.11
    Master of Applied Science
    University of Toronto, Toronto, Canada
    Electrical and Computer Engineering
    • GPA: 3.86/4
  • 2017.09 - 2022.02
    Bachelor of Science
    University of Tehran, Tehran, Iran
    Electrical Engineering, Minor in Computer Engineering
    • GPA: 3.92/4

Skills

Programming
Python
C/C++
Linux Shell Scripting
MATLAB
SQL
R
Verilog
AI Libraries & Frameworks
PyTorch
TensorFlow
Keras
Langchain
LlamaIndex
NLTK
OpenCV
AutoML
Fastai
JAX
Development Tools & Skills
Git
Slurm
Docker
MySQL
AWS
GCP
Vertex AI
BigQuery
Sagemaker
Bedrock
Wandb
HDF5
Cloud Computing

Awards

  • 2022.09.01
    Research Fellowship and Student Fellowship
    University of Toronto
    Awarded research and student fellowships for contributions to academic research.
  • 2022.01.01
    Best Industrial Thesis Award
    University of Tehran, ECE Department
    Recognized for having the best industrial thesis among all undergraduate students graduating from the ECE department.
  • 2020.09.01
    Ranked 1st in Sub-discipline
    University of Tehran
    Ranked 1st in my sub-discipline and 7th out of approximately 115 Electrical Engineering undergraduate students (top 10%).
  • 2017.09.01
    Full Scholarship
    University of Tehran
    Received a full tuition-free scholarship based on academic excellence.
  • 2017.09.01
    Faculty of Engineering Excellence Award
    University of Tehran
    Recognized for academic excellence for five consecutive years and received scholarships from the Supporter Foundation of the University of Tehran.
  • 2010.09.01
    Bronze Medal, Iran's Chess Federation Tournament
    Iran's Chess Federation
    Achieved bronze medal in a national chess competition.

Volunteer

  • Scientific Association, University of Tehran
    Volunteered in multiple events, including the 'How to Apply' symposium and 'Eydaneh' (New Year event).
  • Data Science Winter School
    Contributed to the organization and execution of the Data Science Winter School.
  • Iranian Association at the University of Toronto (UTIRAN)
    Assisted in organizing cultural and networking events to support the Iranian student community at the University of Toronto.
  • IEEE Scientific Association, University of Tehran
    Actively participated in the IEEE University of Tehran Student Branch.

References

Prof. Stark Draper
My amazing supervisor during my master's studies. Professor @ ECE UofT
Shayaan Mehdi
The most supportive manager I ever had! AI Project Management @ the Vector Institute