Trisha Sharma
Analyst
I explore data to spot trends and turn them into useful insights. I build ML/DL models that predict, explain, and simplify complex information so anyone can understand it.
TECH STACK
Hi, I’m Trisha. I work with data, build ML/DL models, and turn complex patterns into simple insights that actually make sense. I enjoy solving problems, experimenting with AI, and creating things that make data feel less intimidating and more useful.
Currently, I'm a final year stundent Manipal University Jaipur, pursuing a degree in Data Science and Engineering. I have maintained a strong academic record, consistently achieving high grades and earning a place on the Dean's List for multiple semesters.
Outside of work, I’m into photography, traveling, and I’m a big fan of movies and series—especially sci-fi adventures and crime mysteries. I also love ramen and anime.
Summer Intern · Ericsson
At Ericsson (Gurgaon), I cut financial variance analysis time by 60% using Python with Azure OpenAI and Azure Storage, automating data retrieval, processing, Excel integration, and NLP for sharper insights. I also engineered NLP solutions for two enterprise clients, including an email summarization system, fine-tuned LLM models, and Spanish-English translation processing, while gaining valuable exposure to corporate culture and working with brilliant minds.
Machine Learning Intern · Cloud Riverdale Pty Limited
At Cloud Riverdale Pty Limited (Australia, Remote), I built Python data pipelines for spam email detection using Hugging Face LLaMA, TensorFlow regression models, and ETL processes for real-time classification. I also integrated a Grok API chatbot into a clothing e-commerce Wix site, applying prompt engineering and data preprocessing techniques that deflected 30% of routine customer inquiries.
Team Lead · Capstone Program by Deloitte
During the Deloitte Capstone Program (Remote), I was selected among the top 20 from 240+ students for a 3-month mentorship, leading Team Data Mavericks on a Generative AI project for automated video monitoring and behavioral assessment. I integrated a YOLO DeepSORT detection model with Streamlit, enabling real-time anomaly alerts and improving incident response times by 30%.
Vice President · Turing Sapiens (IOT Club)
I served as Vice President of Turing Sapiens, the tech club at MUJ, where I organized and hosted various technical events while mentoring juniors. This experience not only allowed me to contribute to the university’s tech community but also strengthened my leadership, collaboration, and teamwork skills.
Gen AI for Video Monitoring and Automated summary and Assessment
Deployed a video surveillance solution using the SPHAR dataset, achieving 95% precision in event classification. Reduced incident response times by 40% for the monitoring team. Built an intelligent pipeline with automatic alerts, detailed incident summaries, and a dashboard interface that identified the top three causes of false alarms.
COVID-19 Detection Using CNN Through Chest X-Ray Images
Built a CNN model with convolutional layers, dropout, and Adam optimizer, achieving 60% accuracy in COVID-19 chest X-ray classification. Enhanced preprocessing with OpenCV, boosting training speed by 30% and improving accuracy by 8%, enabling faster diagnoses.
Adidas Sales Dashboard Using Power BI
Designed a Power BI dashboard for Adidas sales, analyzing over 10,000 data points to uncover 15% revenue growth opportunities and improve regional sales insights.