01
Autonomy Simulation Lab
Interactive autonomy and localization environment combining planning, noisy sensing, nonlinear localization, Kalman filtering, telemetry, and quantitative evaluation.
GNSS-inspired measurementsKalman estimate: tracking
KUSH RISHIMONTRÉAL, CANADA
INTELLIGENCE
UNDER
UNCERTAINTY.
Machine learning, ML systems, PNT, and intelligent sensing. Building research-oriented software for models and systems that must remain reliable when reality stops matching the training set.
BEHAVIOR ATLAS / SYNTHETIC DEMO
ROBUSTNESS0.764
FAILURE RATE23.6%
REGRESSION Δ+4.0 pp
SCENARIO STRESS36
MODEL BEHAVIOR
FAILURE FORENSICS
ROBUST EVALUATION
PHYSICAL-WORLD INTELLIGENCE
00 / CURRENT RESEARCH
RESEARCH / NOVELTY REVIEW / V0.1 DESIGN
MODEL REGRESSION
FORENSICS
What if a model regression could be traced back through training lineage to the change that caused it, then verified with an intervention instead of guessed from correlation?
Enter project →01 / SELECTED SYSTEMS
Explore system →
02
CareBridge Canada
Full-stack product prototype exploring source-grounded, safety-aware AI workflows with React, FastAPI, SQL, automated testing, and CI/CD.
→→
TRACE / GROUNDPASS
claim_01[S1, S3]
claim_02[S2]
unsupportedBLOCKED
02 / RESEARCH DIRECTIONS
03 / TRAJECTORY
FROM MEASUREMENT
TO INTELLIGENCE.
My foundation is in Geomatics Engineering, GNSS/PNT, sensing, and measurement systems. I now work across Python/Linux engineering, data pipelines, validation, and observability while building deeper research and engineering depth in machine learning.
I am particularly interested in model evaluation, ML systems, multimodal intelligence, and how learned systems behave under distribution shift, incomplete information, and real-world uncertainty.
2026—GNSS Analyst · Xona
2025—26Research Assistant · University of Calgary
2022—26BSc Geomatics Engineering · With Distinction