AJ
Calgary, Canada · Open to opportunities

Annette
John

Software & AI Engineer. Researcher. Mountain chaser. Avid reader.

I build AI systems that matter — from safety-aware clinical ML to agentic developer tooling. When I'm not writing code, I'm chasing summits in the Rockies or losing myself in a good book.

AJ

5

Publications

4+

Years building

3

Countries lived

Summits to climb

About

Human first,
engineer second.

I'm a Software & AI Engineer currently pioneering the AI/ML engineering practice at Urban Systems in Calgary. Before that, I spent two years at the University of Calgary doing ML research that genuinely mattered — building systems to help clinicians make better decisions for prostate cancer patients.

Before Canada, I was a software engineer at Credit Suisse in Pune, building the backend systems that move money across borders. I learned a lot about what "reliability" actually means when the stakes are real.

Outside of work, I'm drawn to the mountains — there's something about altitude and effort that resets my thinking better than anything else. I read widely, write occasionally, and believe the best engineers are the ones who stay curious about everything.

M.Sc Computer Science, UCalgaryAWS CertifiedGCP CertifiedBest Thesis Nominee

Career

Where I've worked.

Urban Systems

Software & Machine Learning Developer

Feb 2026 – Present

Calgary, Canada

  • Pioneering the company's first dedicated AI/ML engineering practice — directing AI agents (Claude Code, Cursor, GitHub Copilot) across the full development lifecycle.
  • Building full-stack internal apps, REST APIs, and data pipelines with Python/Django and React; leveraging MCP integrations and agentic patterns to automate workflows.
  • Establishing team-wide standards for agentic development and sharing effective practices across the org.

University of Calgary

Graduate Researcher & Teaching Assistant

Sep 2023 – Dec 2025

Calgary, Canada

  • Built production-oriented ML pipelines for large-scale temporal and multimodal EHR data on Google Cloud.
  • Developed risk prediction, forecasting, and resource optimization models — including mixture-of-experts and container orchestration ML systems.
  • Served as TA for CPSC 331, 433, 471, DATA 221, 601, 604, 605 — supporting instruction, tutorials, and assessment.

Credit Suisse (UBS AG)

Software Engineer

Jul 2021 – Aug 2023

Pune, India

  • Developed backend services (Java, SpringBoot) for international payment systems covering FX, securities, and high-value cross-border transactions per SWIFT/SIC standards.
  • Spearheaded credentials migration to secret management, mitigating $50M+ in compliance risk. Led reconciliation tech stack migration saving ~$3M.
  • Optimized high-contention datastore operations cutting latency by 60%; built internal tool automating SLM process, saving 24 hours/month.

Intellico Labs

Software Engineering Intern

Jun 2020 – Jul 2020

Pune, India

  • Led website development and SEO initiatives; recognized as Best Performer among 25 interns.

Projects

Things I've built.

Master's Thesis

TwinRx

Safety-aware digital twin ML system for inpatient risk prediction and treatment recommendation using large-scale EHR data. Novel mixture-of-experts model for subgroup-aware outcome prediction and causal inference pipelines.

MLCausal InferenceHealthcare AIPyTorchGCP

Azure AI

OracleRAG

Retrieval Augmented Generation system integrating Azure AI Search + Azure OpenAI for factually-grounded answers. 95% accuracy, 90% retrieval precision through modular retrieval/generation design.

RAGLLMsAzure OpenAIVector Search

Predictive Container Orchestration

BiLSTM + Random Forest hybrid for cloud resource prediction, outperforming Kubernetes default scheduling. Reduced average job wait time by 28% via proactive container placement predictions.

BiLSTMKubernetesCloudForecasting

Air Pollution Hotspot Detection

ML-driven geospatial analytics using Sentinel-5P satellite data to detect pollution hotspots and identify pollutant sources. Interactive maps for spatial interpretation.

Deep LearningGeospatialSatellite Data

Face Detection with AWS Rekognition

Cloud-based face detection application using AWS EC2, S3, and Rekognition for image processing and inference.

AWSComputer VisionCloud

YelpCamp

Full-stack campground discovery and booking platform built with the MEAN stack. Passport.js authentication, MongoDB Atlas, deployed to production.

Node.jsExpressMongoDB

Life

Beyond the code.

The things that make me who I am when the terminal is closed.

Currently Reading

The Pragmatic Programmer

David Thomas & Andrew Hunt

65% through

Every chapter feels like a conversation with the engineers I wish I had met earlier. The 'broken windows' metaphor alone is worth the read.

Recently finished

Thinking, Fast and SlowPsychology
The InnovatorsHistory of Tech
Outdoor Adventures
🏔️Hiking

Banff National Park

Plain of Six Glaciers trail at sunrise

🏔️Ice Walking

Lake Louise

Frozen lake in January — surreal silence

🌲Trail Running

Kananaskis Country

Chester Lake loop in autumn colours

🌊Hiking

Johnston Canyon

Upper falls icefall in winter

Stargazing

Jasper National Park

Dark sky preserve, no words do it justice

🏔️Scrambling

Ha Ling Peak

Vertical scramble, 360° views of Canmore

Papers I Found Interesting
Foundation2017

Attention Is All You Need

Vaswani et al.

Still the most elegant architectural decision in modern AI. The simplicity of self-attention as a universal mechanism is genuinely beautiful.

AGI2023

Sparks of Artificial General Intelligence: Early Experiments with GPT-4

Bubeck et al. (Microsoft Research)

The most thoughtful attempt I've seen to grapple seriously with what 'general intelligence' might mean in practice.

Biology + ML2021

AlphaFold: A Solution to a 50-Year-Old Grand Challenge in Biology

Jumper et al. (DeepMind)

Proof that ML can genuinely solve problems that felt intractable. A reminder of why this field matters.

Causal Inference2020

Causal Inference: What If

Hernán & Robins

Made me realize how much 'standard' ML ignores causality. Changed how I frame every modelling problem.

Blog / Ideas / Thoughts
AI & Dev6 minApr 2025

What Building AI Agents Actually Taught Me About Software Design

After months of directing Claude Code, Cursor, and GitHub Copilot across a full dev lifecycle — here's what surprised me most about where humans still matter.

Research Life4 minFeb 2025

The Romanticization of the PhD

Nobody tells you that most of a research degree is quiet, unglamorous effort. And that's actually where the good stuff lives.

Personal5 minNov 2024

On Being an Immigrant in Tech

Moving from Pune to Calgary for a Master's while navigating a new country, a new culture, and a new career — some honest thoughts.

Research

Published work.

A Systematic Review of AI as a Digital Twin for Prostate Cancer Care

2024 · Computer Methods and Programs in Biomedicine, Elsevier

Digital TwinsHealthcare AISurvey

Towards Digital Twins in Prostate Cancer: A Mixture-of-Experts Framework for Multitask Prognostics in Hospital Admissions

2024 · Applied Sciences

Mixture of ExpertsClinical MLPrognostics

Estimating Treatment Effects and Recommending Safe Drug Policies in Prostate Cancer Inpatients: A Causal Inference Approach towards Digital Twins

2024 · Network Modeling Analysis in Health Informatics and Bioinformatics

Causal InferenceTreatment EffectsDrug Policy

Predictive Container Orchestration in the Cloud using AI Techniques

2023 · Computing

CloudKubernetesForecasting

Detection of Air Pollution Hotspots and its Sources using Deep Learning

2022 · Journal of Emerging Technologies and Innovative Research

Deep LearningGeospatialEnvironment

Toolkit

What I work with.

Data & AI

Machine LearningDeep LearningLLMsRAGCausal InferencePyTorchTensorFlowScikit-learnPandasPySpark

AI Dev Tooling

Claude CodeCursorGitHub CopilotMCPAgentic Workflows

Backend & APIs

PythonJavaDjangoSpringBootREST APIsgRPCKafkaMicroservices

Frontend

ReactTypeScriptNext.jsTailwind CSSNode.js

Data Stores

PostgreSQLMySQLMongoDBDynamoDBBigQuery

Cloud & DevOps

GCPAWSAzureDockerKubernetesTerraformJenkinsCI/CD

Contact

Let's connect.

Whether it's a project, a role, a research idea, or just a good book recommendation — I'd love to hear from you.