Jan 2024 – Present · 2 yrs 5 mos · Remote

• Tried my stint at teaching. I was engaged part-time (on weekends) with Great Learning, where I mentored courses on Artificial Intelligence. This role allowed me to share my knowledge and enthusiasm with aspiring data scientists, helped them to build a strong foundation in AI and Machine Learning

Mar 2023 – Mar 2025 · 2 yrs · Bengaluru, Karnataka, India · Hybrid

• Working on uplifting SIEMENS portfolio in Generative AI. Finding & Driving avenues where application of Generative AI can lead to significant productivity improvements, & process efficiency.
• Built Agile Gen AI suite for Jira, Business proposal generator from Business requirement documents, Mr. Agile - An automated daily scrum Bot to make standups even more effective!

Mar 2022 – Feb 2023 · 11 mos · Bengaluru, Karnataka, India · Hybrid

• Automotive parts anomaly detection Platform
• 1. Creation of HLD/LLD documents for the SAAS platform
• 2. Custom unsupervised models trained and Benchmarked NVIDIA Triton server, TAO toolkit, Open Vino
• 3. Creation of Cython / C++ codes for low latency inference ( High resolution images at 11 Fps )
• 4. MLOps practice design which include usage of Mlflow, DVC pipeline
• 5. Process parellalization using multithreading and async functions
• Drone Intelligence Platform
• 1. Automated REST api based object detection training pipeline
• 2. Inference pipeline involving multiple services leveraging RabbitMQ
• 3. Docker image optimization and size reduction for seamless deployment
• 4. Contributed to CI pipeline caching mechanism optimization
• 5. Super tiny object detection model and sliding window inferencing implementation

Mar 2021 – Feb 2022 · 11 mos · Bangalore Urban, Karnataka, India

• Built Data Science platform from scratch. Data Pipeline to Deployment as a Full Stack Data Scientist.
• 1. Product Onboarding pipeline leveraging Labelimg tool
• 2. Multi model Inference using paddleocr, easyocr, kerasocr, natural language processing
• 3. Automated testing using PyTest and Test coverage report generation
• 4. Fastapi REST services creation and Docker containerisation
• 5. Postgres Database creation/migration using Alembic and Flyway
• 6. Logging using customised python logger
• 7. Enhanced python package management using poetry
• 8. Code vulnerability scanning using Veracode
• 9. Integration with Cargill CI/CD pipeline and AWS Sagemaker
• 10. API key based authentication package using FastAPI, compatible with Postgres/ SQL lite DB

Oct 2020 – Mar 2021 · 5 mos

• In this role I worked as a customer face of QpiAI for engagement with customers. I supervised the development of microservices such as Knowledge Discovery, Predictive Maintenance and Model Deployment
• Knowledge Discovery : Worked on single line diagram entities extraction. Digitalizing of printed documents captured from mobile devices. More algorithms to add in this microservice.
• Predictive Maintenance : Basic streaming data monitoring microservice based on Univariate time series forecasting and anomaly detection.
• RUL with ACGAN : Research paper submitted with team for RUL estimations using ACGANs on Spectrogram and Recurrence plots on time series.
• OCR on Drum disks : Automation of production delivery line, captured information of drum disks produced using mounted camera and saved in a database. Took ensembled model approach for meeting very high accuracy criterion. Used CRAFT/EAST and EASY OCR as DL-OCR stack.

Aug 2020 – Oct 2020 · 2 mos

• 1)Created a microservice for QpiAI-Pro AutoML platform called Deploy. That microservice is responsible for deployment and scaling of DL CV models on edge devices and cloud, generated using QpiAI-Pro. Developed Quantization, Pruning, Distillation schemes for speeding up DL CV models and configured Open-Vino and NNCF for deployment on edge devices.
• 2)Worked with team for scaling using automating dockerization of ML/DL models as app using Flask RestAPIs for seamless deployment
• 3)Worked on DL model conversion ONNX ~ MxNET ~ Torch ~ IR for quick model deployment on different platforms.

Jul 2018 – Jul 2020 · 2 yrs · Bengaluru, Karnataka, India

• Delivered AI solutions in Video Analytics, including Object Classification (ResNet, PyTorch) and Object Detection (YOLO, RetinaFace).
• Developed advanced Object Counting systems using image processing (Hough Circles, Blob Detection) and deep learning (TensorFlow, SSD, Faster R-CNN).
• Led 3D Object Reconstruction POC projects using Structure from Motion and Stereo Imaging.
• Engineered Handwritten OCR systems for recognizing mathematical equations, leveraging Keras and TensorFlow.
• Built Predictive Analytics Models for defect detection in digital systems, applying techniques like Logistic Regression, Random Forest, and XGBoost.

May 2017 – Jul 2017 · 2 mos · Bangalore

• Developed a Survellience System to detect objects in a cluttered scene using feature engineering, and finding difference between two images taken from different sources, under different lighting conditions and invariant to rotation and translation. The model was deployed on edge device (Smartphone)
• In depth analysis of various object detection system from Template Matching to SURF/SIFT and various feature matching techniques.

May 2016 – Jun 2016 · 1 mo · Jaipur Area, India

• An Efficient attendance system for large classroom.
• Studied about Viola-Jones algorithm (uses AdaBoost technique of Machine Learning) implemented and tested it in Matlab. Worked on a method of taking attendance which required one-time training and it worked very efficiently and was found reliable.