payam sedighiani

• Set up AWS cloud for collecting data from 240 sensors in Caterpillar engines (CAT). • Implemented InfluxDB for efficient data visualization and extraction. • Designed Co-GNNs, GNNs, LSTM, ARIMA, SARIMA, and FBPROPHET models with 95-99% accuracy forecasts. • Developed anomaly detection models specific to Caterpillar engines. • Created monitoring panels for sensors and anomalies, enabling predictive maintenance. • Retrained BERT model for automatically filling reports using masking

• Developed the first application that combines real-time audio voice cloning and translation functionalities in an audio call • Trained real time zero shot TTS with cloning voice with VITS model for three languages French, English, and Portuguese • Created and worked with multiple speech to text models such as google STT, Speechmatics. • Designed real time STT with 99 percent accuracy. • Designed sentence tokenizer from scratch for unpunctuated texts for multiple languages.

• Worked as Computer vision engineer for hardware and software development. • Designed an I/O board for deploying computer vision models and algorithms for deploying models for securing workers in the coasts while loading and unloading ships and heavy machines. • Designed a 2-Layer PCB using Kicad 5 and kicad 6 and jlcpcb website. • Designed STM 32 microcontroller circuitry, Wiznet w5500, optocouplers and buck converters circuitries. • Designed the board with 8 DOs and 4DIs

• Researched 6 months successfully in the robotics team as a computer vision developer. • Worked on Google Smart Glass using Python programming and OpenCV, dlib, haar_casacade. • Developed Aruco marker detection, Hand detection, Logistical box detection models. • Developed a program in which workers of logistic can be guided to pick and pack items using Google Smart glasses.