Data Engineer at Speechmatics

Integrated 3 new languages (Urdu, Bengali, Swahili) into the ASR pipeline; developed and deployed a Welcome Voice agent for Flow showcased at CES; built an LLM testing pipeline; created interactive dashboards with Plotly Dash for nontechnical stakeholders; designed automated storage reduction scripts for stale experiments; debugged Bash and Python pipelines to resolve CUDA memory bottlenecks, optimize GPU utilization, and improve ASR decoding accuracy; evaluated storage performance (NFS vs S3) and optimized pipelines; developed automated benchmarking to compare ASR accuracy of competitor models on medical datasets across multiple languages; led cross-functional collaboration with PMs and sponsors to drive ASR commercialization and prioritize languages by ROI.

Data Engineer / Developer at Office for National Statistics

Lead developer: Built and led the development of two successful pipelines in PySpark processing UK’s Trade; improved published figures by reflecting additions and removals of unrecorded traded goods; transitioned pipelines from Cloudera (Development) to Data Access Platform (DAP) cloud environment; developed ETL pipeline to transfer and transform data to centralized storage using Hive (Data Warehousing) and SQL; maintained best practices through comprehensive testing, RAP principles, documentation and more; regularly presented technical information to non-technical end users for project milestones (discovery, alpha, and beta gates); translated user requirements into technical specifications; demonstrated leadership by line managing and upskilling two junior developers.