Backend Engineer - Air.ai (2023-09 - 2024-09)
The world's first self-learning, multi-modal conversational AI that can sales/CS calls using phone.
• Developed the Django backend for Generating Voice according to customer's request and maintained upgrades combined FastAPI.
• Optimized the open-source TTS model, BARK, with TensorRT and ONNX, reduced the bot's response delay time from 2.5s to 1.5s.
• Designed the implementation of continuous integration and continuous deployment (CI/CD) pipelines using Kubernetes to automate and enhance code deployment workflows.
• Increased the bot's smartness through advanced algorithms and effective prompt engineering techniques.
Backend Engineer - HayStack (2019-9 - 2023-08)
HayStack reinvented the intranet for the future of work, with a focus on seamless integrations, powerful search capabilities, and consumer-grade design.
• As a Backend Engineer at talkhealth.ai website, creted strategic plans, suggesting innovative solutions, and managing end-to-end product lifecycles.
• Pioneered in the development of a dynamic trading bot using python and leveraging RNN, LSTM, and Transformer models for accurate time series prediction.
• Implemented Bitcoin Price Prediction utilizing Neural Networks, several specialized algorithms, and an innovative approach to news analytics. Successfully predicted considerable price trends, leading to higher profit margins.
• Designed the backend of the home.speaksynk.com project, which supports voice interpretation for over 10 languages, and reduced model delay time by fine-tuned the Bark Model.
Full stack Engineer - Devmatics (2017-05 – 2019-07)
Devmatics is a professional software development company that provides custom software, mobile apps, integration, and workflow automation.
• Designed the Remote Teaching System equipped with a facial recognition system using AI and various teaching support functions. Developed and Deployed the Whole project using FastAPI, Github and Django for 3 months.
• Upgraded backend functionalities for Fluent.AI (https://fluent.ai) by implementing a scalable voice model that accommodates various lengths: this led to a 25% increase in user retention and positive feedback from over 500 users.
• Participated in Voice Bot Development using Eleven labs, Deep Gram, DID and another Neural Network.
• Acquired knowledge in AI-related technologies: BERT, T5, BARK, GPT, LLM and Vector Databases such as Pinecone, Chroma DB, Milvus.