Omid

Senior Machine Learning Engineer with 5+ years of experience architecting AI systems end-to-end. Proven track record in developing scalable, production-grade RAG systems, LLM-powered multi-agent frameworks, and cloud-native ML pipelines. Expertise spans Python, PyTorch, Hugging Face, vector search, prompt orchestration, and cloud deployment (AWS, Azure). Experienced in multimodal AI, knowledge graph extraction, and autonomous agents.

Founding Machine Learning Engineer - EVOLO.Ai - Toronto, Canada

(2024-04)

• Built the entire AI system from scratch, including multi-agent system and multimodal retrieval pipelines.
• Developed a fully automated ETL URL-to-RAG pipeline: web scraping, raw storage, chunking, embedding, and vector DB ingestion.
• Engineered a multimodal RAG system integrating image scraping, captioning, metadata tagging, and Titan Multimodal embeddings.
• Designed dynamic multi-agent orchestration using LangChain and AutoGen with tool routing logic.
• Created a knowledge graph extractor with LangGraph to enrich the RAG system with hierarchical entity relations.
• Implemented summarization and email follow-up tools leveraging RAG outputs for business-facing use cases.
• Led end-to-end infrastructure on AWS (S3, Lambda, EventBridge, Bedrock), containerized with Docker.
• Conducted prompt engineering and built evaluation pipelines for continuous tuning and improvement.
• Created LLM evaluation pipelines to benchmark performance using semantic similarity and faithfulness metrics.
• Worked on designing and implementing Multi-Tenancy and Data Residency strategies to support secure, scalable AI systems.

Data Engineer - HOLMETRICS - Calgary, Canada

(2023-06 - 2023-07)

• Actively worked with AWS and its services like EC2 instances, S3 buckets, RDS, Redshift, Lambda, Glue, and Sagemaker.
• Experienced in working on data warehouses and data lakes, utilizing MySQL and PostgreSQL through Datagrip to effectively manage big data.
• Executed end-to-end ETL pipelines, encompassing data collection from APIs using Postman and Requests, followed by processing and storage in relational databases.
• Utilization of Cron Jobs and EventBridge for pipeline scheduling, as well as SQS and load balancers to ensure scalability.

Machine Learning Developer - ALTAML - Calgary, Canada

(2023-01 - 2023-04)

• Designed a recommender system for Suncor, enabling the company to provide safety recommendations for workers to prevent incidents.
• Worked with Azure computing instances, utilizing Azure tools such as Form Recognizer and gaining a strong understanding of cloud computing.
• Used Git for version control in the team, gained a high level of expertise in the tool, performed PR reviews, and learned to resolve merge conflicts.
• Learned how to work with data pipelines, using Make and config files, and got familiar with code bases and workflows.

Machine Learning Developer - University of Calgary - Calgary, Canada

(2022-09 - 2024)

• Developed expertise in 3D computer vision, including using ML techniques and models to analyze and interpret complex 3D medical data.
• Implemented an unsupervised domain adaptation technique with Unet to generate accurate segmentations for unlabeled data.
• Worked with professional Linux commands, bash scripts, and the Slurm Workload Manager to code with computer clusters and Tensor Cores.
• Applied knowledge distillation to train a cost-effective student neural network from a large teacher model, reducing computational cost while maintaining performance.

Co-Founder and Machine Learning Developer - Artificial Intelligence and Robotics Center (AIRCENTER) - Tehran, Iran

(2020-08 - 2022-08)

• Demonstrated leadership skills by managing a team of five for two years, overseeing projects' development, and conflict resolution.
• Successfully delivered over 10 freelance ML projects to clients, with details of selected projects included in the project section of the resume.
• Instructed ML and DL courses to more than 1200 students as part of our educational program, of which I was the instructor.

Machine Learning Developer - Iran's National Elites Foundation - Tehran, Iran

(2020-09 - 2021-08)

• Developed a multi-stage deep system consisting of four stages of cascaded detection, including overtaking lane detection, HGV detection, license plate detection, and character recognition.
• Worked on the London Traffic Dataset and employed the YOLOv5 model for object detection, OCR for character recognition, Hough transform, and Canny for line detection.

Machine Learning Developer - IUST - Tehran, Iran

(2019-09 - 2021-09)

• Developed an intelligent modular vision-based system for environment perception.
• Implemented real-time semantic segmentation with the SGDepth model.
• Performed monocular depth estimation with a non-stereo camera based on deep learning techniques.
• Worked on real-time line detection with the PINet model.
• Implemented image processing techniques like convex hull, morphological erosion, and dilation for noise canceling and ROI creation.

MSc - Computer and Software Engineering - University of Calgary (2022 - 2024)

BSc - Computer Engineering - Iran University of Science and Technology (IUST) (2017 - 2021)