Tech Lead with 15+ years shipping AI-augmented enterprise platforms at regulated-industry scale - insurance, healthcare, financial services, higher education. Four-plus years into a multi-year DB2 -> FAST migration at Prudential - 10M+ policies, 5,000 contracts, 10-20 engineers across 4 domain squads. Shipped a RAG assistant on Azure OpenAI GPT-4 + LangChain adopted by 30+ analysts: retrieval MRR above 0.85, hallucination rate under 3%, citation-grounded. AWS Certified Solutions Architect - Associate.
Overview
15
15
years of professional experience
Work History
Tech Lead
Acro Systems Inc - Prudential Financial
, Remote
10.2021 - Current
Lead 10-20 engineers across four domain squads on a multi-year DB2 -> FAST migration of 10M+ policies and 5,000+ contracts.
Contract-by-contract reconciliation diffs gate every cut-over window; 99.9% read availability preserved across releases.
Shipped a RAG assistant on Azure OpenAI GPT-4 + LangChain over ETL lineage and COBOL metadata, adopted by 30+ analysts; citation-grounded answers cut investigation from ~4 hours to ~8 minutes, with a human-in-loop eval harness holding retrieval MRR above 0.85 and hallucination rate under 3% before every prompt-template release.
Own the DB2 -> Python extract -> S3 -> AWS Glue -> FAST pipeline moving 150K+ CSV files (20-30 per contract across 5,000 contracts), with schema-drift checks and per-contract audit trails.
Modeled the Contract domain as an event-sourced write side on AWS EventBridge with CQRS read projections; analyst dashboards hit purpose-built read models at p95 under 120ms - roughly 10x faster than the mainframe-backed path they replaced.
Built a Backend-for-Frontend layer that collapses 12-15 downstream calls into a single Contract-shaped payload per view, insulating React remotes from system-of-record churn during the FAST cut-over.
Stood up a Micro-Frontend platform on Webpack 5 Module Federation - four domain teams ship React remotes independently into one host shell; release cadence per domain moved from quarterly to weekly.
Hardened customer identity with CIAM (OAuth 2.0 / JWT / OIDC) across legacy and modernized UIs, unifying sign-on across the migration boundary.
Rebuilt delivery on GitHub Actions + Jenkins with contract tests, automated canary deploys, and observability gates; deployment lead time dropped from days to under 2 hours; change-failure rate under 5%.
Designed cloud-native integration platform connecting 3 enterprise systems (Unifier, KFS, Jaggaer) via 8 Spring Boot microservices on Red Hat OpenShift, processing 5,000+ daily transactions.
Implemented event-driven architecture with RabbitMQ handling 10,000+ messages/day, achieving sub-second synchronization of purchase orders, requisitions, and invoices with 99.95% delivery reliability.
Developed 4 Oracle APEX web applications for payroll and parking services, automating deduction workflows that eliminated 25 hours/week of manual processing for 3,000+ employees.
Deployed 8 containerized services on OpenShift using ConfigMap, Secrets, and Service Registry, reducing deployment cycles from 2 days to 3 hours with zero-downtime releases.
Tech: Java 8, Spring Boot, Red Hat OpenShift, Oracle APEX/ORDS, RabbitMQ, Oracle Exadata
Lead Software Engineer
Washington University in St. Louis
St. Louis, MO
02.2018 - 01.2020
Led development of Bio-Specimen Management System using Java EE, Grails, and Spring, managing 200,000+ clinical specimen records across 15 research studies with 100% HIPAA compliance.
Designed 8 Pentaho ETL pipelines processing 500,000+ clinical records monthly, reducing data preparation time from 3 days to 4 hours for observational study analysis.
Built automated reporting platform generating 50+ weekly clinical reports via BIRT, reducing manual report creation from 4 hours to 15 minutes per report.
Managed CI/CD pipelines with Jenkins and Ansible deploying to 6 RHEL environments, cutting release cycles from 2 weeks to 2 days with 95% automated test coverage.
Developed HIPAA-compliant MDM platform on IBM InfoSphere processing 8M+ patient records from 12 source systems, achieving 99.97% uptime for mission-critical healthcare operations.
Built high-throughput batch processing systems using Spring Batch, processing 1.5M+ records nightly and reducing batch completion time by 50% through chunk-based optimization.
AI Gateway Platform, Built centralized API layer for LLM access (Azure OpenAI, AWS Bedrock). Implemented rate limiting, cost tracking, and audit logging. Standardized LLM usage across multiple applications.
Zero-Downtime Data Migration & Reconciliation Engine, Orchestrates large-scale data migrations with zero downtime. Automated validation and reconciliation with configurable rules. Built with Python, AWS Glue, S3, and Apache Spark.
Cost-Aware Observability Platform, Built monitoring system using OpenTelemetry, Prometheus, and Grafana. Correlated logs, metrics, and traces for system-level insights. Intelligent cost optimization across infrastructure.
ShopKrisha.com, Full-stack e-commerce platform with product catalog, cart, and checkout. Built with React, Next.js, Node.js, TypeScript.