About Me

Hi! I’m a full-stack Data Scientist and AI/ML Lead with 6+ years of experience designing, deploying, and scaling production-grade machine learning systems across industries — including automotive, utilities, finance, and healthcare.

With a PhD in Structural Engineering and deep expertise in machine learning, NLP, deep learning, and MLOps, I specialize in transforming raw data into enterprise-ready AI solutions. I’ve led development on platforms like Quality.AI at Ford, delivering measurable results like a 25% reduction in engineering issue resolution time and $5M in annual cost savings.

I bring a rare blend of engineering, data science, and product strategy. I’ve served as a Product Manager, shipping AI-driven features across global teams, translating user needs into roadmap priorities, and aligning technical delivery with business value.

I’m passionate about AI for impact — whether it’s accelerating diagnostics, improving customer experience, or enabling responsible AI adoption. My work integrates modern cloud environments (AWS, GCP, Azure), distributed data pipelines (Spark, PySpark), and deployment tools (Docker, GitHub Actions, Streamlit) to bridge the gap between idea and impact.

Additionally, I work on my personal data science projects. You can learn more about them by reading my blog. I’m always happy to chat about data science and related topics. If you’d like to contact me, please use the contact form, or drop me a line via LinkedIn or email.

My Toolbox

Soft Skills

Technical Leadership, Cross-Functional Collaboration, Mentoring, Communication

Programming

Python, SQL, R, MATLAB, Bash, Git

Machine Learning

LLM Fine-tuning, NLP, Time Series Forecasting, Deep Learning, Probabilistic Modeling, Feature Engineering

Data Visualization and Communication

Power BI, Tableau, Looker

Testimonials

Parviz Soroushian - CEO & Founder

In addition to the broad scientific value of Dr. Harsini’s research, manifested in the citation of his work and its inclusion in eminent journals and professional conferences, the progress in development of innovative products is priceless in terms of its commercial applications.

Fred Faridazar - Federal Project Manager

Dr. Harsini developed an analysis procedure that directly evaluates the effect of the model input parameters of pavement structures on the pavement performance using Monte Carlo simulation and Artificial Neural Network modeling. This is a robust method to conduct sensitivity analysis for a given pavement structure. This method would enable the designer to put more efforts on the accurate determination of highly significant variables.

Anagi Balachandra - R&D Director

The work of Dr. Harsini’s research on the novel non-destructive system and accompanying data acquisition and software system were applauded enthusiastically by the Technical Panel of Experts assembled by FHWA. Dr. Harsini contributed to this work on different levels. He had direct contribution in preparation of the proposal which led to $750,000 funding by FHWA. During the implementation phase, he played a key role in developing and testing of the system. Some of his findings on this subject have already been published in ASCE Journal of Materials in Civil Engineering.

Neeraj Buch - Professor and Chairperson

Dr. Harsini has authored a patent on the use of recycled glass in concrete that has been widely used by campus of Michigan State and city of East Lansing. This work has reflected in local and national news outlets and social media, and has attracted several commercial groups offering collaboration and strong support.

Venkatesh Kodur - Distinguished Professor

Dr. Harsini has good grasp of fundamental engineering principals and good research abilities. Dr. Harsini’s expertise gained at MSU, is valuable to real-world applications.

Syed Waqar Haider - Associate Professor

Dr. Harsini, has initiated the use of Neural Network and Latin Hypercube Sampling in the sensitivity analysis of the performance predication models in pavement engineering to identify the important factors for pavement analysis and design. Dr. Harsini has also developed algorithms that could be used for the optimization and sensitivity problems in other fields of science and engineering when numerous input variables are involved with prediction models. His work is unique and important in assisting practitioners to understand more complex systems in an exclusive way.

Behrouz Shafei - Assistant Professor

Dr. Harsini has already accomplished a great deal as a young engineer/scientist, and has advanced the field of engineering in terms of nanotechnology, computational mathematics science and other topics which have exciting application potential in any number of areas. It is a testament to his impressive record that he has already published several papers for respected journals and conferences.