Sorry, unable to follow Ian. Please try again.
I am an Senior AI Engineer at the Defence Science and Technology Agency (DSTA), where I specialize in bridging the gap between cutting-edge research and mission-critical applications. My work focuses on harnessing State-of-the-Art (SOTA) NLP and speech technologies to build resilient, scalable solutions for the defense sector. Currently, I enjoy deploying end-to-end containerized speech-based services, ensuring that high-performance AI models are integrated seamlessly into secure environments using containerized solutions. Beyond deployment, I am deeply involved in fine-tuning and evaluating NLP models to meet specific business requirements, ensuring that our AI is not just powerful, but purposeful. I am driven by the rapid evolution of speech and language modeling and dedicated to evaluating and implementing the next generation of SOTA models to maintain a strategic edge in AI innovation.
- Research and implementation of state-of-the-art speaker identification and diarization AI models - Employed Python packages like Pytorch, Numpy, Librosa, Gradio etc. - Implemented deep learning models like ResNets, Time-Delayed Neural Networks, Convolutional Neural Networks etc. - Finetuned and utilised pre-trained models from Nvidia NeMo, Pyannote, HuggingFace, Github
- Utilised Python APIs like PySpark, MLFlow, Pandas, Sklearn etc. for big data (a few hundred thousand data points) analysis. - Implemented Machine Learning algorithms for anomaly detection in production data. Algorithms such as Kmeans Clustering, Isolation Forest, One-Class SVM, Neural Networks etc. - Established a data pipeline with Microsoft Azure Data Bricks to run production data through Machine Learning models and output them into a DeltaLake database. - Created visualization of big data through Python code using DataBricks Notebooks.
- Taught students key computing concepts - Organised and managed computing camps - Gained and demonstrated technical expertise and people management