Mar 2023 – May 2025 · 2 yrs 2 mos · Freiburg, Baden-Württemberg, Germany · On-site

• 1. Developed a RAG-based chatbot for the web platform used by ~50 employees and ~150 municipalities in Germany.
• 2. Leveraged OpenAI Assistants for contextual conversations with images and speech-to-text with the bot.
• 3. Developed Django REST APIs with PostgreSQL database. Gathered all the feature requirements and specifications.
• 4. Collaborated with a team comprising 1 Director, 1 Solution Architect, 1 Backend Engineer.

Dec 2022 – May 2025 · 2 yrs 5 mos · Freiburg, Baden-Württemberg, Germany · Hybrid

• 1. Assisted topics - MLPs, Backpropagation, Optimization, Regularization, CNNs, RNNs, Transformers, GANs & Diffusion.
• 2. Managed the course material, graded assignments, prepared exam questions and tutored ~100 Master’s students during Winter Semesters 2023/24 and 2024/25. Mentored 3 new Student TAs - Ashwin K., Shaza K., Enes H.
• 3. Developed an automated closed-captioning pipeline for the Deep Learning and AutoML lectures, leveraging FFMPEG and OpenAI’s Whisper. Contributed to the course launch on YouTube (youtube.com/@automl-freiburg-edu).
• 4. Collaborated with a team comprising 2 Professors, 3 Postdoctoral Researchers, 3 PhD candidates, and 8 student assistants.

Sep 2020 – Apr 2022 · 1 yr 7 mos

• 1. Product Management for Deductions AI - Validity Prediction (class-imbalance problem), Claims Merge, Reason Code.
• 2. DVP average recall of ~92% across 20+ clients | ~75% Claims Merge rate | ~80% Reason Coding rate.
• 3. DVP ~8% average improvement via business feature engineering - time-series rollback feature capturing recency.
• 4. FTE & OPEX savings for 5 live active clients across NA, EU, APAC (Kellogg's, Kraft Heinz, NBTY, Smithfield, Hormel).
• 5. Charted the business-specific domain playbook for lateral hires.
• 6. Additional: Spearheaded the HighWay to HighRadius Internship Program. Streamlined the hiring process through ML automation, reducing efforts from ~160 to ~30 hrs (~5.33x reduction).
• 7. Collaborated with: 2 VPs, 1 Director Products, 1 PM, 2 Team Leads, 5 POD Leads.

Jul 2019 – Aug 2020 · 1 yr 1 mo

• 1. Worked on the Deduction Validity Predictor use case, focusing on building and deploying machine learning models.
• 2. Developed ML models using Python, NumPy, Pandas, Matplotlib, and Scikit-learn, ensuring high accuracy and performance.
• 3. Deployed models through RIVANA, an in-house deployment engine.
• 4. In the last 4 months, I slowly transitioned from ML Engineering to ML Product Management.

May 2019 – Jun 2019 · 1 mo

• 1. Contributed to the development of an AI-enabled FinTech B2B cloud application.
• 2. Utilized Java Servlet Programming, ExtJS, SQL, and Python to build and optimize key functionalities.

Aug 2018 – Sep 2018 · 1 mo

• Worked on the official website for the company. Below are the tools and technologies which were used while working on the project,
• 1. Python & Django - Python was used to write the backend for the entire web application. This is a framework of Python which can be used to build Web Applications using the concept of REST APIs.
• 2. HTML5, CSS3, Bootstrap, Java Script - All these technologies were used to build the User Interface i.e the frontend of the entire web application.

May 2018 – Jun 2018 · 1 mo · Mumbai Area, India

• Worked on a project titled Big Data Inventory Management System. The business impact of the web application includes,
• 1. Engaging day to day trade for several goods and services. From handling Purchase Requisitions, Sales Orders, Invoice Data, Monthly Sale and so on.
• 2. Able to handle multiple types of users like Manufacturers, Distributors, and Employees.
• 3. The system could send automated emails to the Manufacturers for Manufacturing next set of goods and services using the concept of Inventory Precision and Predictive Production.
• 4. Also, able to show user-specific analytics in the analytics dashboard.
• Below are the tools and technologies which were used while working on the project,
• 1. Python & Django - Python was used to write the backend for the entire web application. This is a framework of Python which can be used to build Web Applications using the concept of REST APIs.
• 2. HTML5, CSS3, Bootstrap, Java Script - All these technologies were used to build the User Interface i.e the frontend of the entire web application.
• 3. SQLite - This was used as the database for storing all the information.
• 4. Bokeh - Python-based library for building visualisations and analytics dashboards on User Interfaces.