Hi, I'm Samridhi, a machine learning engineer. I build end-to-end ML systems, from model training and evaluation to cloud deployment and CI/CD. My work spans multimodal LLMs, agentic systems, RAG pipelines, and evaluations. Most recently, I built an AI scribe used by 200+ physicians across 105 facilities. 4+ years across ML/AI and software engineering.
/work
Machine Learning Consultant
2025 – present
Alberta Health Services
Leading ML/LLM feature design with clinical teams. Building agent-based evaluation
pipelines for AI-generated clinical summaries.
Machine Learning Engineer
2023 – present
Faculty of Medicine & Dentistry, University of Alberta
Designed a multimodal AI scribe (Whisper + LLaMA 3.3 70B) serving 200+ physicians.
Built RAG agents for clinical knowledge. Deployed on GCP and AWS.
Machine Learning Consultant
2022 – 2023
Alberta Machine Intelligence Institute (Amii)
Consulted with 4 AI startups on ML integration, predictive analytics roadmaps,
and lightweight MLOps patterns.
AI Mentor & Teaching Fellow
2022 – 2024
AI4Good Lab
Mentored 36 members from underrepresented groups through full AI project lifecycles. Led a team to win the AI4Good Lab Accelerator Award 2023, securing $2,000+ in funding. Facilitated ML and project management sessions.
Graduate Research Assistant
2022 – 2023
University of Alberta
Developed QDTL, a transfer learning method achieving 17.7% reduction in prediction error for medical applications. Led interdisciplinary teams to 97.6% accuracy in organ transplant predictions.
Software Engineer
2019 – 2021
Samsung Research & Development Institute
Built automation tools and conformance test suites for global Samsung TV product lines.
Reduced QA time by 14%.
/projects
Berta AI Scribe
· Python, Next.js
Multimodal AI medical scribe combining speech recognition and large language models
for clinical documentation at scale.
Medical Predictive Modelling
· Python
Master's thesis on applying transfer learning techniques to medical predictive modelling.
InterStellar
· Python
Cryptocurrency price prediction app using LSTM neural networks, with a Lumen dashboard for visualizing Stellar price forecasts.
Footfall
· R, Shiny
Retail footfall analysis dashboard built with R Shiny. Tracks customer traffic, purchases, and conversions to generate marketing and store design recommendations.
/research
An open-source AI scribe platform for clinical documentation, deployed across Alberta Health Services
serving 198 emergency physicians at 105 facilities. Combines automatic speech recognition with large
language models while maintaining data within secure institutional infrastructure, achieving a 70-95%
cost reduction compared to commercial alternatives.
Quality-Diversity Transfer Learning (QDTL)
· OpenReview, 2024
A novel transfer learning approach based on neuroevolution for small-dataset medical
prediction problems. Combines quality-diversity optimization, architecture search,
and conceptual expansion to outperform standard transfer learning baselines.
More on Google Scholar
/writing
Co-authored with Louis Bouchard and Ben Lorica. A deep dive into AI deep research
tools, their comparative landscape, and how they're reshaping knowledge work.
Gradient Flow, Apr 2025
/reading
I Am That
now reading
The Heart of the Buddha's Teaching
Meditations
The Power of Your Subconscious Mind
More on Goodreads