Ahmet Canik

• Developing and deploying production-ready Real-Time Speech Recognition models for Pega Voice AI, contributing to its impact on changing the agent work experience through speech-to-text and real-time AI. • Performing comprehensive testing, verification, and optimization of trained AI models, ensuring consistent high-quality performance. • Integrating and deploying trained models into production by developing optimized code, streamlining AI integration.

• Developed and deployed an automated ML training pipeline for Real-Time Speech Recognition, taking full ownership from data collection and cleaning to model training, fine-tuning, and testing. Utilized various speech recognition frameworks, primarily in Python and C++. • Designed and implemented automated benchmark and regression tests, guaranteeing the quality of newly trained models. • Engaged in R&D on machine learning applications in Speech Recognition, NLP, Language Models, and Text Summarization.

• Trained, debugged and tuned real-time Automatic Speech Recognition model with Kaldi framework. • Applied post processing NLP methods to improve ASR accuracy. • Implemented real time voice metrics (pitch, volume) to existing ASR system.

• Built Speech Analytics solutions for call centers, from network packet capturing to web front-end. Used libpcap for packet capturing, kafka for transmission packets between components, Spring Boot for backend, ElasticSearch for storing and searching, Kibana for monitoring, and Angular for frontend. • Built Call Recorder for Call Centers. Network packets are captured and parsed, then saved in any desired format. Deep knowledge of VOIP signaling and data protocols such as SIP, H323, Alcatel NOE and RTP.

Oracle 11g & 12c RAC installation, management, maintenance, backup, performance optimization Built a media server solution that holds media files (audio, video, document) using Oracle RAC with ACFS and SecureFiles. Developed HTTP Proxy with Java 8. Data synchronization and migration with Oracle Data Integrator 12c. Deployed ModSecurity and ModProxy to secure internal application servers. OWASP ModSecurity rule set applied.

Developed web crawler for internet data harvesting and a query application to query the data obtained. This system is used for multiple purposes for both specific data and general internet open-source data. Coded with C#. Developed a streaming media system. Created a framework to use as an infrastructure for software projects. Framework enables new data bindings and query&report screens with zero code change. Coded with C# and PL/SQL.

Passed all courses with grade A. Artifical Intelligence Machine Learning Data Mining