Senior Software Engineer at Walmart Global Tech (2022-09 – Present)
Led backend reliability across global commerce workflows, owning the end-to-end data lifecycle from source events to analytics and downstream decisions. Designed resilient data ingestion patterns for complex financial sources and ensured system reliability and scalability under fast-paced demand. Built core platform infrastructure that supported ML-ready pipelines for anomaly detection use cases and automated financial insights reporting at production scale.
- Architected the infrastructure for a resilient data ingestion and normalization pipelines framework, improving processing consistency across high-volume streams and lowering downstream incident rates.
- Built feature extraction pipelines feeding ML-ready pipelines, enabling anomaly detection signals to appear in production workflows within 11 days for operational triage.
- Productionized ML models by wrapping inference services with authentication checks, improving system reliability and scalability for near real-time decisioning.
- Developed vendor-agnostic abstraction layers for API/SDK consumption, improving integration stability across multiple internal clients and keeping schema changes manageable.
- Ensured system reliability and scalability by designing resilient retry and backpressure logic, reducing ingestion lag during peak traffic by 37%.
- Ingested banking data event feeds via secure ingestion jobs, improving data freshness for automated financial insights and enabling consistent downstream categorization.
- Design and operate resilient data ingestion with canonical models alignment, ensuring canonical models remained stable despite upstream field volatility.
- Shaping the core product by owning ingestion-to-insights delivery, coordinating webhooks and rate limiting policies to keep downstream systems responsive.
Senior Software Engineer at Parallel Finance (2022-03 – 2022-09)
Delivered greenfield infrastructure for Web3 lending workflows, shaping the core product around reliable data feeds and ML-ready pipelines. Built resilient data ingestion for complex financial sources, with careful Postgres schema design and SQL migrations to support evolving analytics. Developed income categorization logic and anomaly detection heuristics to power automated financial insights.
- Owned the data lifecycle for new reporting features, aligning data ingestion and data normalization pipelines so downstream systems consumed consistent records.
- Designed Postgres schema design for canonical models, enabling safe SQL migrations and faster iteration on model inputs without breaking consumers.
- Built income categorization pipelines using feature extraction pipelines, improving classification coverage by 28% across self-contained transaction sets.
- Developed and productionized ML models as repeatable services, improving anomaly detection signal latency to 2.4 seconds in staging tests.
- Implemented authentication and webhooks integration for partner systems, reducing manual reconciliation work by 19% while preserving system reliability.
- Ingested tax transcripts-like accounting artifacts from external sources, standardizing shapes through data normalization pipelines for dependable analysis.
- Consumed external APIs/SDKs through resilient client layers, ensuring vendor-agnostic abstraction layers handled retries and edge-case payloads.
- Ensure rate limiting protections on critical endpoints, preventing overload and improving scalability during bursts from automated financial insights dashboards.
Software Engineer at ServiceNow (2021-01 – 2022-03)
Built and operated platform services that supported enterprise workflow automation, focusing on resilient data ingestion and consistent delivery into downstream processing. Developed TypeScript services with careful authentication and rate limiting to keep integrations reliable. Designed production-grade data pipelines that supported feature extraction pipelines and canonical models, with strong operational practices for system reliability.
- Designed API gateway rules with rate limiting and authentication, ensuring system reliability and scalability for high-throughput automation workflows.
- Built TypeScript ingestion services for data ingestion and data normalization pipelines, improving correctness of transformed records across 14 job types.
- Developed feature extraction pipelines that produced model-ready datasets, enabling canonical models alignment for analytics and operational reporting.
- Architected SQL migrations for incremental schema evolution, reducing deployment risk by 34% during peak release windows.
- Implemented webhooks for downstream event consumption, improving integration responsiveness while preserving end-to-end data lifecycle integrity.
- Consumed internal APIs/SDKs using vendor-agnostic abstraction layers, reducing coupling and keeping ingestion stable during upstream changes.
Software Engineer at Intuit (2017-06 – 2021-01)
Worked on financial software used by individuals and accountants, delivering ingestion and analytics infrastructure tied to tax transcripts and accounting platforms. Built ML-ready pipelines and feature extraction pipelines to support income categorization and anomaly detection workflows. Owned production delivery and productionize ML models work to support automated financial insights and dependable user outcomes.
- Built resilient data ingestion pipelines for banking data feeds, improving end-to-end data lifecycle completeness by 22% across core product flows.
- Developed Postgres schema design for accounting platforms reporting entities, supporting SQL migrations and keeping data transformations consistent under schema churn.
- Ingested tax transcripts and standardized fields via data normalization pipelines, enabling accurate income categorization at scale.
- Developed feature extraction pipelines for ML-ready pipelines, improving the quality of model inputs for automated financial insights by 16%.
- Productionized ML models with authentication-gated access patterns, ensuring system reliability and scalability for time-sensitive inference workloads.
