Elham M.
Senior Applied Scientist @Dialpad
Signup · Get unlimited contacts
WORK HISTORY
Senior Applied Scientist @Dialpad
Ottawa, ON, CA
Leading the development of Agent Builder, an internal LLM-driven tool that enables automated workflow generation for customer support teams- Built a multi-turn embedding model for contextual one-step search, cutting query latency and inference cost- Designed and deployed a real-time PII redaction model, improving transcript precision while maintaining recall- Contributed to training and optimization of Dialpad’s in-house LLM (DPGPT) to support longer contexts and reduce deployment bottlenecks- Collaborated cross-functionally to troubleshoot infrastructure and language expansion challenges, enhancing deployment workflows.
EDUCATION
University of Tehran
Bachelor of Arts (B.A.), French Language and Literature
Concordia University
Master of Science (M.Sc.), Computer Science
National Organization for Development of Exceptional Talents (NODET)
High School, Physics and Mathematics
University of Tehran
Master of Science (M.Sc.), Computational Linguistics
University of Tehran
Bachelor of Arts (B.A.), English Language and Literature
ABOUT ELHAM M.
I\'m a Senior Applied Scientist specializing in Natural Language Processing with a track record of delivering end-to-end AI systems from research to deployment. My work bridges deep learning innovation and real-world scalability, building products that are both intelligent and production-ready.At Dialpad, I\'ve led several AI core initiatives including Agent Builder, a generative AI platform that uses LLMs to automate customer interactions by creating agentic workflows. I\'ve also developed multi-turn embedding models for contextual retrieval, real-time redaction systems, and contributed to evaluation pipelines that ensure LLM quality across migrations and language expansion efforts.My technical background spans model fine-tuning, model optimization, and inference efficiency using tools like PyTorch, Hugging Face, ONNX, and Vertex AI on GCP and AWS. I\'m equally comfortable solving infrastructure bottlenecks, designing model evaluation frameworks, or aligning AI outputs with business goals.I\'m passionate about building reliable, explainable, and efficient language systems that make AI truly useful, not just powerful.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.