Experience

Splashgain Technologies

Aug 2023 – Mar 2024

Pune, India

Conversational SystemsAngularJSSQLHighcharts

Built and deployed a production conversational system for Eklavvya and e-Pravesh, automating support for 500+ global customers and serving 5,000+ concurrent users on a platform processing 30M+ assessments in 15+ countries.
Redesigned AngularJS components for Eklavvya, improving data-loading workflows and UI responsiveness by 35% during peak examination traffic.
Developed SQL stored procedures for backend retrieval and reporting, improving database access efficiency for assessment and administrative operations.
Built Highcharts monitoring dashboards for grades, comparative rankings, and student-performance analytics across more than 30M assessments.

Splashgain Technologies

Jul 2022 – Jul 2023

Pune, India

TensorFlowEfficientDet-D0Computer VisionChatbots

Developed a real-time computer-vision pipeline processing webcam streams at 20+ FPS to detect unauthorized phones and earphones without disrupting live examinations.
Fine-tuned EfficientDet-D0 with TensorFlow and custom annotated datasets, validating detection across varied lighting conditions and webcam hardware.
Identified recurring support queries and introduced an automated chatbot workflow, reducing manual responses and laying the foundation for broader conversational-support automation.

Mahindra and Mahindra

Aug 2021 – Sep 2021

Nashik, India

PythonPandasNumPyMongoDBLSTM

Processed a 4GB CNC telemetry dataset using chunk-based Pandas ingestion into MongoDB, overcoming memory constraints to analyze more than 112K servo-speed observations.
Built a time-series preprocessing pipeline with Pandas, NumPy, MongoDB, and linear interpolation, converting irregular machine signals into cycle-aligned sequences and handling missing values.
Developed an unsupervised LSTM autoencoder, training on 41,639 sequences and evaluating 9,432 test samples; achieved 0.4705 maximum MAE loss and detected 48 anomalous sequences.