Software Engineer Machine Learning

Surya Prakash Sharma

Software Engineer/AI ML Engineer

+918********* # *********************@*****.*** ï Linkedln § GithHub Work Experience

Holmium Technologies Pvt. Ltd., Noida Present - Jan 2024 Software Engineer

• Constructed interactive web apps using HTML, CSS, JS (frontend) and Flask/FastAPI with PostgreSQL (back- end), integrating Celery with Redis for background tasks and improving load times by 60% via performance testing.

• Enhanced HTTP communication through RESTful APIs for ESP-based microdevices, achieving a 60% reduction in page load time.

• Optimised PyQt software with Google Cloud API and service account authentication, enabling server-to-server communication. Reduced load times by 50% and increased user satisfaction by 40%.

• Orchestrated biweekly code reviews for a 5-member team, identifying 15+ critical bugs per session. Skills

Coursework: Machine Learning, Deep Learning, Artificial Neural Networks, NLP, Computer Vision. Programming Languages: HTML, CSS, SQL, JavaScript, Python, C++. Frameworks: FastAPI, Flask, OpenAI’s GPT, Streamlit, PyQt, Django. Libraries: Pytorch, Numpy, Pandas, Matplotlib, Scikit-Learn, TensorFlow, Keras, NLTK. Mathematics: Linear Algebra, Statistics, Probability, Calculus. Developer Tools: Git, GitHub, MySQL, VSCode, Jupyter Notebook, Google Colab. Projects

Football Analysis Link

• Engineered a football analysis tool leveraging YOLOv5 and tracking algorithms for seamless real-time recognition of 22+ players and the ball with over 90% detection accuracy.

• Implemented camera movement estimation and perspective transformation algorithms, enabling speed and distance calculations with an error margin under 5%.

• Applied unsupervised ML (KMeans) for automated player colour classification and OpenCV-based ball path interpolation, improving visual clarity by 40% during analysis.

• Technical Skills Applied: Supervision, Scikit-learn(Kmeans), OpenCV, YOLOv5, Python, Pandas, Numpy. Sentiment Analysis Link

• Analysed 10,000+ book sentences for sentiment (positive/negative/neutral) with 85%+ accuracy.

• Applied NLP techniques like tokenisation, word frequency, and stopword removal.

• Improved preprocessing and regex logic, boosting categorization consistency by 30%.

• Engineering Stack: NLP, Regular Expression, NLTK, Python. Historical Weather Data API using Flask Link

• Constructed a Flask-based REST API to serve historical weather data with support for dynamic queries.

• Leveraged Pandas to process and retrieve weather data from over 50,000+ records with high accuracy.

• Resolved data consistency and optimized API response time by 40%, enabling smooth integration with external systems.

• Utilized Technologies: flask, Pandas, Python.

Drowsiness Detection Link

• Developed a real-time drowsiness detection system using facial landmarks and the EAR, achieving over 90% detection accuracy.

• Integrated an audio alert mechanism with Pygame, reducing user response time to drowsiness by 2–3 seconds.

• Handled image processing and detection challenges using OpenCV and dlib, excel frame processing speed by 35%.

• Key Technologies: OpenCV, Pygame, dlib, sciPy, Python. Education

Shoolini University (Solan, HP, India)

Department: B.Tech Computer Science and Engineering CGPA : 8.67/10 August 2019 - June 2023

Contact this candidate