Hello, I'm Suraj Patil.

I formulated the approach for adoption of Artificial-Intelligence for the traditional manufacturing setup and I am passionate to work on crucial data analytics projects. My core skill sets involve Gen AI, deep learning model making and adaptive systems solution development. I am also a PhD research scholar at Department of Technology, Savitribai Phule Pune University, Pune, India. My research interests are in the fields of Biometrics, Signal Processing, Computer Vision and Financial Analysis. I am actively pursuing developmental projects with concrete research components which can be made into production in an end to end fashion using deep learning models, Hybrid Mobile App frameworks like Flutter and Serverless architectures.

Work Experience:

Publications:

7) Patil, Suraj, Manik Hendre, and Aditya Abhyankar. "Fingerprint reconstruction from minutia using wave atom basis function." Multimedia Tools and Applications (2024): 1-26.
6) S. Patil, M. Hendre and A. Abhyankar, "Fingerprint Quality Metric Fusion Framework," 2023 IEEE Pune Section International Conference (PuneCon), Pune, India, 2023, pp. 1-4, doi: 10.1109/PuneCon58714.2023.10450009.
5) S. Patil, M. Hendre and A. Abhyankar, "Alignment Free Fingerprint Template Security Using Minutia Neighbors," 2023 IEEE International Carnahan Conference on Security Technology (ICCST), Pune, India, 2023, pp. 1-4, doi: 10.1109/ICCST59048.2023.10474241.
4) Hendre, M., Patil, S. & Abhyankar, A. Biometric recognition robust to partial and poor quality fingerprints using distinctive region adaptive SIFT keypoint fusion. Multimed Tools Appl 81, 17483–17507 (2022).
3) Hendre, M., Patil, S. & Abhyankar, A. Directional filter bank-based fingerprint image quality. Pattern Anal Applic 25, 379–393 (2022).
2) Hendre, Manik, Suraj Patil, and Aditya Abhyankar. "Utility of quality metrics in partial fingerprint recognition." International Journal of Computing and Digital Systems (2021).
1) S. P. Patil and V. D. Nagrale, " Robust adaptive beamforming with positive semi-definite constraint using single variable minimization ," 2015 International Conference on Industrial Instrumentation and Control (ICIC), Pune, 2015, pp. 1587-1590, doi: 10.1109/IIC.2015.7151003.

Research and Development Projects:

Jul 2020 - Oct 2020 | Drawing League App | DL Startup
Problem Statement required development of image scoring ML models, testing and deploying into production. 5K+ images in database were used for training the deep learning model. Mobile app is in deployment phase, soon to be live. Carried out all the phases in SDLC for full stack development.
Tools: Cloud Firestore, Heroku, Git, Flutter, Dart, Python, TensorFlow, Keras, Flask, JavaScript, HTML, CSS

Nov 2019 - Jun 2020 | Stockassis App | Stockassis
Stockassis was designed to be the world's #1 estimator with self evaluation, that is stock centric and also provides self evaluatory measure for all the popular BSE30 stocks. Weighted average of regression methods was used for prediction based on different indicators of stock market. App was published on Google Play Store and was monetized using ads. Carried out all the phases in SDLC for full stack development. Dedicated Dashboard for the app was made using Tableau.
Tools: Flutter, Dart, Backendless REST-API, Firebase, Admob, Python, Scikit-Learn, Tableau, BeautifulSoup

Oct 2017 - Present | PhD Research Scholar | DoT, SPPU, India
Thesis: Towards Building Fingerprint Framework For Enhanced and Secured Matching
Proposed an iterative finite time Hill-Climbing algorithm based fingerprint Attack on fingerprint recognition system. Worked on U-net based Deep Learning Segmentation algorithms for fingerprint attack. Current work is focussed on the phase based AM-FM fingerprint model to increase minutiae feature representation for higher matching performance.
Tools: Python, MatLab, TensorFlow, Keras, Scikit-Learn, Pandas

Sep 2016 - Oct 2020 | Project Assistant | RUSA - Centre For Intelligent Systems | DoT, SPPU, India
Sponsored Project: Liveness Based Anti-Spoofing Measures. Funding of 0.25 Cr- RUSA Grant, Govt. of India.
Micro ATM Product launched in market. Order of 50 supplied to Lokmangalam, order of 70 supplied to Cloud Moyo. In a team of 4, carried out preparation of fingerprint specimen, integration of scanner SDK's with networking frameworks, 3D design and printing of Micro ATM prototype. Fingerprint Liveness detection based on image classification using statistical learning (SVM and K-NN) and deep learning techniques.
Tools: Java, Python, TensorFlow, Keras, OpenCV, Scikit-Learn, Numpy, SciPy, Raspberry-Pi, Stratasys- uPrint SE Plus 3D printer

Jul 2014 - Jun 2015 | Master's Student, SPPU
Master's Thesis: Robust adaptive beam-forming for general-rank signal model with positive semi-definite (PSD) constraint for optimal solution
We propose a two step closed form solution of the formulated robust adaptive beamforming problem, wherein a new single variable minimization problem is constructed. Result of this minimization is used to solve the robust adaptive beamforming problem. Using Lagrange’s Multiplier Method, a closed form solution for the robust adaptive beam-forming problem was achieved. Simulation results verify the improvement in the performance by the proposed method over the current robust adaptive beamforming methods for general-rank signal model.
Tools: MatLab, Python, SciLab.

Jul 2012 - Jun 2013 | B.E. E&TC SPPU
B.E. Project: Subcutaneous vein detection system using infra red imaging technology.
Proposed system revealed the veins on the surface of the hand. It captured the vein images using infrared technique, processed these images, and displayed them on screen. It formed a low cost and non-invasive scanning procedure.
Tools: MatLab, MultiSim, Infrared camera, TV Tuner card.

Courses and Projects:
Oct 2020 | Udemy Course
Course: Python & Machine Learning for Financial Analysis.
Lectures: 131 | Video: 23 Hours | Certificate
Projects [ Github Link]:
1) Stock Price Predictions Using Times Series Analysis
2) Bank Customers Segmentation Using Credit Card Data For Marketting Data Analytics
3) Stocks Sentiment Analysis For Twitter Data
Algorithms: Ridge Regression, K-Means, PCA, LSTM
Tools: TensorFlow, Keras, NLTK, Gensim, Scikit-Learn, Numpy, Seaborn, Plotly, Matplotlib, Pandas

Jun 2020 | Linux Academy
Course: AWS Essentials 2019
Lectures: 78 | Video: 7.5 Hours

Feb 2018 | Deeplearning.ai
Course: Convolutional Neural Networks
Lectures: 42 | Video: 6 Hours

Mar 2016 - May 2016 | PhD CourseWork, DoT, SPPU
Project: Electroencephalogram (EEG) based vehicle directional control.
Tools: MatLab, Neurosky EEG kit, Raspberry-Pi, Putty, DC motors.