C*****I
About Candidate
Backend and distributed systems engineer with experience building large-scale, low-latency platforms in adtech environments. I work on high-throughput backend services and real-time decisioning systems operating under strict latency and reliability constraints.
My work focuses on system optimization and scalability, including concurrency, traffic management, and integrating machine-learning models into production systems. I’m particularly interested in improving the efficiency and resilience of complex platforms through thoughtful design and performance-oriented engineering.
I collaborate closely with SRE, data, and ML teams to ensure backend architectures remain stable and reliable during traffic spikes and infrastructure degradation.
Technically, I work primarily with C#, Java, Python and Scala, alongside distributed data pipelines and containerized platforms. I’m motivated by problems at the intersection of scale, latency, and system reliability.
Core skills:
– Distributed Backend Systems (Low-latency, High-throughput)
– Java, Python & C#/.NET in Production
– Kubernetes & Containerized Platforms
– Real-time Data Pipelines & Streaming Systems
– Production ML Serving (LLM-based classification, Triton)
– Observability, Performance Tuning, Reliability
Nationality
Location
Education
Work & Experience
• Worked on real-time infrastructure serving 700M+ daily users, processing billions of events with millisecond-level latency and handling up to 300M QPS at peak. • PCA Modernization (PCAv2) — Modernized 20+ legacy services and improved observability (Python, Scala); integrated an LLM-trained classification model served on NVIDIA Triton, powering the classification of 500M+ webpages/day. • Bid Request Logging Optimization — Implemented conditional filtering in a core C#/.NET Core ingestion component, reducing dataset size by 40% and saving $1M/year in infra cost while preserving analytical value. • Traffic Shaping for RTB (key TLA) Stability — Built an online profitability scorer and SLA-based rejection controller in C#/.NET Core; combined real-time load signals with 144h historical analysis to drop low-value traffic under bursts or DC degradation, protecting SLA and margin. • Arbitrage (key TLA) Parallelisation — Parallelized a latency-critical arbitration workflow in C#/.NET Core using async/Task, Parallel.F
• Distributed data pipelines on Hadoop/Spark/YARN/Kubernetes supporting large-scale analytics across finance and strategy domains. • Ownership of ingestion and processing modules within the enterprise data platform, including schema design and data quality enforcement. • Spark/Hive performance optimisation and cluster monitoring (Ambari) improving stability and reducing execution time. • Exploratory analysis and modelling of large semi-structured datasets (JSON, CSV) using SparkSQL.
Machine learning models for anomaly detection using Python/MLlib, including feature extraction via clustering. • Visualization dashboards delivering production-quality metrics and insights to manufacturing teams. • Analytics methodology workshops on anomaly detection, visualization, and advanced statistical techniques.
