My technical journey in AI/ML has developed through a combination of academic study and hands-on projects that bridge mechanical engineering with applied machine learning. As a B.Tech Mechanical Engineering student at NIT Trichy with a CGPA of 8.95, I have pursued practical applications that address real infrastructure and logistics challenges. This statement outlines my experiences to date, identifies remaining knowledge gaps, and explains why my background positions me well for the MLSS programme.
Projects and Practical Experience
During my summer internship at Chipmonk Solutions in 2025, I contributed to an AI-powered Intelligent Parking System. The work involved constructing computer vision pipelines that employed CNNs and Faster R-CNN models alongside OpenCV for vehicle detection and Automatic Number Plate Recognition. I also helped implement time-series forecasting using LSTM networks, integrated with Kafka for event processing and MongoDB for data storage, to support occupancy monitoring and utilisation analytics delivered through React.js dashboards. This experience demonstrated how deep learning models can be combined with streaming data architectures to produce operational tools.
Separately, I developed an AI-driven Supply Chain Management platform using XGBoost for demand forecasting. The pipeline incorporated lag features, rolling statistics and seasonal decomposition to refine inventory predictions, followed by simulation routines for safety-stock calculation and reorder optimisation. Visual analytics were provided via Streamlit. In a further remote research project on Automated Damage Detection for Electrical Towers, I processed 3D LiDAR point clouds with PointNet++ architectures and fused outputs with YOLO-based image detection for cross-modal inspection workflows. These projects illustrate a pattern of applying supervised learning techniques to structured sensor and image data while managing end-to-end pipelines.
Knowledge Gaps and Programme Alignment
Despite these applied experiences, my understanding of theoretical foundations remains limited, particularly in probabilistic modelling, optimisation theory and the statistical guarantees that underpin modern architectures. Exposure to reinforcement learning, unsupervised representation learning and the ethical deployment of large-scale systems has been minimal. MLSS would address these gaps through structured instruction that moves beyond implementation toward principled design choices. My mechanical engineering training supplies domain knowledge in physical systems, which complements algorithmic study and supports the interpretation of model behaviour in real environments. The combination of documented project outcomes, strong academic record and sustained curiosity about scalable learning methods indicates readiness for the programme’s demands and potential to contribute actively to cohort discussions.
Conclusion
In summary, the progression from computer-vision pipelines to multi-modal LiDAR analysis reflects consistent technical growth. The remaining theoretical and methodological gaps align directly with MLSS offerings, while prior results and interdisciplinary perspective strengthen my suitability. Participation would consolidate existing skills into a more rigorous framework.
References
- Qi, C.R., Yi, L., Su, H. and Guibas, L.J. (2017) PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space. Advances in Neural Information Processing Systems.
