Pauroosh Kaushal
Languages spoken: Hindi  [Excellent]
English  [Excellent]

Summary

Edit Summary
<<
Professional Summary

 

Highly skilled Corporate Trainer with a PhD in AI/ML and over 10 years of experience in data analytics, machine learning, and deep learning. Proficient in developing and implementing end-to-end machine learning workflows, including data preprocessing, feature engineering, model development, deployment, and optimization. Adept at delivering tailored training solutions across industries such as marketing, finance, and retail, while mentoring doctoral, postgraduate, and graduate students to enhance research capabilities and maximize performance. Expertise in Python, R, Azure Cloud, and Generative AI models, with a strong focus on data analysis, visualization, and actionable insights. Skilled in handling large structured and unstructured datasets and creating no-code applications to drive innovation and informed decision-making.

 

---

 

Areas of Expertise

 

- Data Analytics: Providing data-driven insights for business problem-solving

- Machine Learning & Deep Learning: Building and deploying advanced models

- Generative AI: Deploying LLM based solutions to cater diverse requirements of clients.

- Data Visualization: Expertise in creating impactful visualizations for corporate reporting

- Machine Learning Model Deployment: Designing and implementing robust and scalable machine learning model deployment on Azure cloud and on-premise platforms

- Generative AI Applications: No-code solutions for AI-driven projects

- Training & Development: Conducting workshops and training for corporate clients and academic institutions

 

---

 

Technical Skills

 

- Languages & Tools:  

  - R, RStudio, Python, Microsoft Azure, Jupyter Notebook, PyCharm, Google Colab, GitHub, MATLAB, VSCode

- Python Packages:  

  - NumPy, Pandas, Scikit-learn, TensorFlow, Keras, Matplotlib, Seaborn

- R Packages:  

  - dplyr, ggplot2, stringr, caret, shiny, RMarkdown, plotly

 

---

 

Professional Experience

 

Koenig Solutions Pvt Ltd

*Corporate Trainer (July 2023 – Present)*  

- Deliver specialized training sessions on data analytics, machine learning, and deep learning to corporate clients.

- Develop custom training materials tailored to client requirements across sectors like marketing, finance, and retail.

- Provide hands-on sessions on advanced data analytics tools, R, Python, and Azure.

 

---

 

Key Projects

- LLM-Integrated Chatbot for Order Query Resolution: Implemented a LangChain-based chatbot integrated with external APIs to automate customer support for online purchase order inquiries. This solution accurately retrieves and presents order details from the database, potentially reducing customer support costs and improving response times.

- Diabetes Prediction System using End-to-End Machine Learning Workflow: developed and deployed an end-to-end machine learning pipeline for diabetes prediction, leveraging Azure Container Instances and a CI/CD pipeline with GitHub Actions for automated deployment. This solution, incorporating data preprocessing, feature engineering, and model optimization, has the potential to improve early diagnosis and personalized treatment plans in healthcare settings.

- Electronic tongue system for beverage and pharmaceutics analysis: Developed an electronic tongue system for beverage analysis, utilizing Wavelet Transform for feature extraction, and computed ML taste models to predict shelf life, ensure quality control, authenticate products, identify regional differentiation and taste masking of pharmaceutics, achieving a 20-45% improvement in quantitative analysis.

- Gait Disorder Prediction System: Developed a multi-class classification model for gait disorder prediction using public datasets and a Random Forest classifier, achieving a 14% improvement in accuracy. This approach has potential applications in healthcare for early diagnosis and personalized treatment of gait disorders.

- Equity and Option Portfolio Management System: Developed a real-time trading strategy using SmartAPI and moving average crossover, achieving a 22% annual ROI. The system also incorporated hypothesis testing for stock return forecasting and Value at Risk (Var) analysis.

---


 

Awards & Certifications

 

- Microsoft Azure AI Engineer

- Microsoft Azure Data Scientist

- GATE 2010: Qualified in Instrumentation Engineering with AIR-929.

- Device Diagnostic Award: Awarded by BETIC, IIT Bombay in 2021 for contributions to medical device diagnostics.

 

---

 

Education & Certifications

 

- PhD in Instrumentation Engineering

 

 

 

Associated with Koenig since July-2023.


Experience

Edit Experience

Project Experience

Multi-class pulmonary disease detection tool for Indian patient (2 year)

  • Collected Pneumonia, Tuberculosis and COVID-19 X-Ray images from public dataset for training and 82 clinical data for validation, image data preparation
  • Implemented image pre-processing and used LBP and GLCM methods for extraction of texture features on Google Colab.
  • Deployed transfer learning based deep learning for pulmonary disease detection with accuracy of 94% on Indian patient.
  • Used Tensorflow and Keras for model development.

Equity and Option Portfolio management (1 year)

  • Developed trading strategy for Nifty 50 data using moving average crossover to achieve ROI of over 22% per annum
  • Performed EDA to predict value at risk (VAR) for portfolio risk management. Forcasting stock returns using hypothesis testing methods to make trading decision.
  • Data visualization to analyze portfolio return, comparison with peer mutual funds

Electronic tongue system for beverage and pharmaceutical sample analysis (4 years)

  • Developed electronic tongue system for data generation. Data optimization, pre-processing and feature extraction using Wavelet Transform
  • Developed CNN models for multi-classification of Indian wine, project supported by National Institute of Grape Research
  • Computed taste model and predicted sweetness in pharmaceutical samples in relation to human panel method. Improved quantitative analysis of beverages by 20- 45%

 

User Identification System using gait analysis (1 year)

  • Collected public gait dataset to develop user authentication system.
  • Data cleaning, pre-processing and deployed random forest classifier, achieving accuracy of 95%

 

Trainings delivered for:

   
LTIMindtreeAlaska(gov)
Angola LNGApple
ASICAustralian Government Department of Defence (Defence, Australia)
BlackwoodsBolton Clarke (Healthcare, Australia)
CapgeminiCognizant Technology Solutions
CRDB Bankcrdbank.co.tz (BFSI, Tanzania)
DellFujitsu
Giesecke+Devrient | G+DH&M
HCLHCL TECH
HPEInfosys
International SOS (Healthcare, Singapore) KPMG US
Maldives monetary authorityMCIT
Microsoft CorporationNovo Nordisk (Pharmaceuticals, India)
Optum (Healthcare, United States) PWC Canada
RMIT University (education, Australia) SABIC (Oil & Gas, Saudi Arabia)
ServiceNowTahakom
Tata Consultancy Services (IT, India) Telstra
The University of SydneyTourism Australia (travel/tourism, Australia)
UBX Tanzania Ltd (IT, Tanzania) Visa
Wipro

Preview Past Training of Pauroosh :

CourseRecording Link

Webinars Delivered

WebinarView Recording
Azure ML: Prompt Flow Testing using Azure OpenAIhttps://youtu.be/Ow-m5E4tp3Q
Developing Generative AI Applications with Gemini and Streamlithttps://youtu.be/skD4wU5yr6o
Exploring Capabilities of Open Source LLMshttps://youtu.be/qignt-CYIUs
From Data to Deployment - Building Machine Learning & AI Solutions with Microsoft Azurehttps://youtu.be/58OTpsMHecY
Leveraging Artificial Intelligence and Data Science with CertNexushttps://koenigsolutionsltd-my.sharepoint.com/:v:/g/personal/info1_koenig-solutions_com/ERc1MkFj0blBim2yDV6rgQABlDKG4R8RSm0UM0BGqwbnWw?nav=eyJyZWZlcnJhbEluZm8iOnsicmVmZXJyYWxBcHAiOiJPbmVEcml2ZUZvckJ1c2luZXNzIiwicmVmZXJyYWxBcHBQbGF0Zm9ybSI6IldlYiIsInJlZmVycmFsTW9kZSI6InZpZXciLCJyZWZlcnJhbFZpZXciOiJNeUZpbGVzTGlua0RpcmVjdCJ9fQ&e=yLS0fH
Use prompt flow testing for evaluating Azure OpenAI modelhttps://youtu.be/Ns4XgCty_Zk

Recent Feedback for Pauroosh :

Student NameFeedback
arjun kulkarni1. Very Knowledgeable and able to answers all questions promptly 2. Very patient and teaching in a good pace 3. Good Verbal skills and interpersonal skills 4. Good Training skills and ensuring the students have understood.
BruceThis feedback should be be considered intoxicating because it highlights multiple positive aspects—clear presentation, helpful insights, professionalism during training, and deep knowledge delivered during training. The feedback leaves participants with a highly positive impression of the trainer.
EdenPauroosh is very thorough with the materials, he always provides examples rather than just reading off the slides, which is great for visual learners. Thank you for leading the training sessions! (especially noting the time zone difference).
JoanPauroosh, your training sessions were truly inspiring! Your deep knowledge, kindness, and dedication made learning enjoyable and impactful. You have a gift for teaching, and your positive energy motivated me to delve deeper into Generative AI. Keep up the fantastic work—you're making a real difference in your students' lives!
NicholasPauroosh was excellent and did a great job of explaining some of the hidden working of ML algorithms and terminology, and was always willing to further explain anything. I also enjoyed how he managed the course strucutre to be in a more logical sequence.
PaulAll participants rated the delivery as 5 (Excellent) and shared positive comments. Notably, the first comment reflects a particularly strong feedback of the delivery. Given the uniformly high ratings and enthusiastic feedback, this can be considered intoxicating feedback and indicative of a highly impactful batch delivery.
Sai KamVery detailed explanations and thoroughly goes through all topics. Also, the willingness to begin at early hours in the morning is commendable.
SergeyHe was very profficient in explaining complex concepts of generative models. He brought a lot of concepts to my mind. And the one which i really liked is ability to use pdf documents context for providing answers .
SharonPauroosh was excellent, he showed a high level of expertise and patience throughout the sessions. He simplified complex topics, presenting them in a way that made the learning experience enjoyable and effective. He was also accommodating and friendly, creating a welcoming environment that encouraged open communication and made it easy for everyone to engage and grasp the content with confidence.
UmeshGreat knowledge on the topic and nicely navigating through the required topics on time and his detailing on clarifications is really excellent.