Communicated with different departments to understand their requirements for ETL pipelines.Built new ETL pipelines using Python and modified existing ones according to requirements.Resolved issues with existing ETL pipelines.Reviewed and merged pull requests for pipeline changes.Handled access permissions for new developers and team members.Added new tables and data sources to existing ETL pipelines.

Developed a comprehensive application for counting people and other objects.Implemented advanced features for detecting the age and gender of individuals.Tracked the movement of individuals within the application.Integrated with InfluxDB to send statistical data to a time-series database.Displayed collected statistics on a Grafana dashboard using Flux QL.Deployed updates and changes to all Raspberry Pi devices.

Led the development of a machine learning pipeline for a major client in the insurance sector. This involved gaining a deep understanding of the client's specific use cases through comprehensive discussions with business stakeholders. Implemented strategies for data collection and cleaning, and executed a variety of database queries based on the specific use case. Developed an Exploratory Data Analysis (EDA) pipeline to visualize the dataset, enabling the identification and removal of redundant features. Authored a comprehensive proposal for a motion detection system, detailing its architecture, advantages, potential drawbacks, and cost estimates.

In this role, I spearheaded the development and implementation of audio recognition software, specifically designed for small IoT devices. The initial phase involved training the software to recognize keywords in voice-activated systems, akin to the "Hey Google" and "Hey Siri" functionalities.I then conducted a comprehensive evaluation of the trained model using multiple evaluation matrices. This was performed on audio samples with diverse accents and noise conditions, ensuring the model's robustness and versatility.Once the evaluation was complete, I optimized and quantified the model. The final step was to convert the model to other frameworks, making it suitable for deployment on a wide range of IoT devices.To ensure the model's performance in different environments, In addition to previous augmentation techniques such as time stretch, and white noise addition, I incorporated augmentation techniques such as speech reverberation, echo, and pitch shifting. This enriched the training and testing dataset, enhancing the model's adaptability.I also conducted several experiments to understand how reducing the training dataset would impact the model's performance. This led to insights on the minimum number of audio required to transfer the model to other keywords using transfer learning techniques. These experiences have significantly contributed to my expertise in audio recognition software development.

- Training and fine-tuning machine learning models- Deploy ML and DL models on the web, android, and Desktops also converting between different frameworks- More than 9 projects of training and deploying Machine Models on web and mobile applicationsMore than 9 projects of training and deploying Machine Models on web and mobile applications

- Enhanced the performance and efficiency of existing data pipelines utilizing PySpark.- Conducted in-depth research on specific use cases, leading to tailored modifications in data transformations within the pipeline to meet evolving requirements.- Identified and resolved production environment errors in the ETL pipeline, ensuring smooth data flow and minimizing downtime.

- I utilized Alteryx to build ETL jobs and leveraged SQL to derive valuable data insights.- I effectively communicated complex findings and insights to non-technical team members in an understandable manner.- I adapted Informatica jobs to accommodate new requirements.- I employed Python to create ETL jobs, extracting data from various servers and loading it into the Hive database following the transformation process.- I assisted with scheduled ETL jobs, handled errors, and communicated any issues to the relevant teams.

- Served as a developer for a thesis project titled "Prediction of Future Location based on Past Trajectories".- Responsible for the implementation of all computer code related to the research.- Developed code for various stages of the research, including data training, testing, and evaluation.- Contributed to the research by providing technical expertise and coding solutions.- Collaborated closely with the research team to understand requirements and provide appropriate coding solutions

- Contributed with a team of 4 members to make a full-stackweb project.- Developed excellent communication skills through the opensharing of ideas and insights with the team.- Contributed to project success via collaboration and by being astrong team player.