May 2025 – Present · 1 yr 1 mo

• Re-engineered legacy data pipelines — updated business logic, added retry mechanisms, and integrated AWS S3 + SNS for real-time failure alerting.
• Built a db-agnostic, fully automated recon pipeline validating replication across 2,000–3,000 tables; supports full, CDC, sample, and custom modes — orchestrated end-to-end via AWS Step Functions with zero manual intervention. It consists of various Data quality checks like primary keys check , null check , column datatype mismatch check.
• Extended the pipeline with a config-driven design (partition count, bucket count, threshold, and more) via config.json, making it portable across any database with zero code changes.
• Built an automated autofix layer to detect and remediate replication anomalies — Duplicate PKs, missing/extra records, and missing columns — without manual triage.
• Cut validation runtime by ~95% (3–4 days → 2–4 hours) on billion-row tables via partition-wise count(*), Spark optimisations (bucketing, repartitioning), and early-exit on primary check failures, mismatch in column numbers, any technical failure like connection timeout.
• Led Redshift Migration Task Planning for QLik Replicate for 2,000–3,000 LUW tables — analysed 6–8 months of CDC data to define ingestion strategy and prioritise tables.
• Architected a 5-stack Qlik Replicate ingestion framework with failure recovery, log stream resilience, and high-availability reload strategies.

Feb 2025 – Apr 2025 · 2 mos

• Successfully graduated from Hashedin University, an intensive training program by Hashedin by deloitte, focussed on full stack development and real-world software engineering practices. Gained hands-on experience with Spring Boot , Angular , .Net , GenAi and Data Engineering ,while building scalable, high-performance applications and working in Agile, collaborative environments.