Madhumita Venkataraman

Data Scientist with expertise in machine learning, deep learning, and data engineering. Successfully designed and implemented scalable AI solutions in various sectors, including environmental mapping and financial analytics. Proficient in Python, R, and TensorFlow, with a strong background in developing ML models and optimizing big data workflows. Recognized for innovative approaches and ability to integrate technical skills with user-focused design.

• I engineered a massively parallel data analytics pipeline capable of processing 400 GB of complex, structured data using Python with MPI4PY on the Spartan high-performance computing cluster to optimize execution, I developed a custom chunking algorithm that distributed data based on byte size, maximizing computational efficiency across core. The solution was rigorously benchmarked on 1-node, 8-core, and 2-node, 8-core environments, and evaluated using Amdahl’s Law to measure scalability. This project achieved complex analytics results in just 5–7 minutes, far outperforming conventional runtimes, and earned me a perfect academic score, making me one of the few students to attain this distinction.
• I designed and implemented a multimodal deep learning framework to track and predict stock prices by fusing historical market trends with real-time news sentiment, and geopolitical and socioeconomic factors. The project leveraged LSTM neural networks for time series forecasting and NLP models to extract sentiment insights from news articles, by correlating qualitative news patterns with quantitative financial indicators, the model offered robust predictive accuracy and actionable insights. This work highlighted my capability to handle multisource data fusion, apply advanced deep learning architectures, and deliver intelligent forecasting systems applicable to real-world financial decision-making.
• Conceptualized and prototyped an innovative device to automate the extraction of Neera (coconut tree sap) a nutrient-rich natural extract shown in multiple studies to support recovery in stage-4 cancer patients. Conducted extensive customer validation and field research, engaging with traditional harvesters to deeply understand inefficiencies and risks in the manual extraction process. The final prototype significantly reduced physical labor, improved safety, and increased yield consistency by introducing automation and optimized suction mechanisms. The project blended human-centered design with engineering precision, addressing a socially impactful agricultural problem.
• Led the development of a machine learning model to assist in real-time cardiac monitoring for Indian Air Force pilots, using ECG signal data from wearable flight suits. Transformed graphical ECG waveforms into structured numerical datasets and collaborated closely with leading cardiologists to extract and mathematically model critical heart-health indicators. Designed a democratic ensemble model capable of predicting 15 distinct heart disorders, achieving an overall accuracy of 87%. The project demonstrated deep interdisciplinary integration combining biomedical signal processing, predictive modeling, and high-stakes operational use cases in aerospace health systems.
• Selected as Team Lead role for a national level competition called the Smart India Hackathon - 2020. The team designed a mobile app to detect early onset of Alzheimer's in various patients. My role in the project was designing ML algorithm using Cobra in addition to overall Team lead activities. Aspire to make app free so as to reach even most remote parts of India.
• Competed and presented in National Institute of Technology (NIT), Trichy for a Festember National Level Hackathon where college team and worked on developing a black box can be fitted into cabs. Black box will succour for women at times of any assault attempted by cab drivers. It was developed in a vision to make night work life for women a bit safer.
• Paper presentation on Big data and AI usage in Banking and Financial services was presented to jury from MNCs, Tamil Nadu.