Karthick T. Sharma
ML Engineer · Healthcare AI · MPhil at QUT
I’m a machine learning engineer in healthcare. I like owning a model from raw data to production, and the part I enjoy most is inference: making models fast enough to run inside a clinical workflow.
At Neuralgap I deployed a TensorRT-optimised 3D CNN-Transformer for cardiac video that runs 140× faster than the legacy system, and built multimodal PyTorch models and Flask APIs for a clinical decision support platform. Earlier, at NTU, I deployed a semi-supervised anomaly detector at a steel manufacturer, and at SUTD I built a streaming sentiment system on Flink and Kafka.
My first-author papers are in the International Journal of Medical Informatics and Applied Soft Computing (both Q1) and at ISBI. I’ve also co-authored papers at two CORE A* conferences, EMNLP and ICDE.
I’m doing an MPhil at QUT on predicting outcomes after children’s heart surgery, and I’m looking for a full-time ML engineering role. I’m based in Brisbane with full work rights in Australia.
I’ve been at a computer since I was five, mostly playing games. I still play, and I’d rather explore the whole map than rush the main quest. I also read philosophy and love cats. Unlike AI, they understand your instructions perfectly. They just don’t care.
கற்றது கைமண் அளவு கல்லாதது உலகளவு
What we have learned is a handful of sand; what we have not learned is as vast as the world.
— Avvaiyar
selected publications
- IJMI
- ISBI
- ASOC
- ICDEScalable Machine Learning for Real-Time Fault Diagnosis in Industrial IoT Cooling Roller SystemsIn IEEE International Conference on Data Engineering (ICDE), 2025
- EMNLPSentiStream: A Co-Training Framework for Adaptive Online Sentiment Analysis in Evolving Data StreamsIn Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023