💼 Career
Experience
Backend, distributed systems, and cloud roles with real impact and metrics.
- 2025-09 — PresentMeta (Contract)Data AnalystTexas, United States
- Collaborate with ML engineers to improve training data quality for Meta's AI and LLM products, reducing evaluation errors by 30%.
- Build Python automation and validation pipelines that process more than 1M annotations per day for production model training.
- Design evaluation frameworks to detect edge cases and bias in model outputs and datasets, improving real world model performance.
- Develop and maintain data quality dashboards and metrics used by more than 20 ML engineers to guide model iteration and deployment.
- Create synthetic data generation workflows to improve coverage of underrepresented scenarios in training datasets.
- Implement active learning feedback loops that connect model predictions with human annotations to optimize retraining.
PythonSQLPandasLangChainData PipelinesInternal ML Tools - 2022-01 — 2023-07Inncircles (Full-time)Software EngineerHyderabad, India
- Designed Spring Boot microservices for contractor matching, increasing user engagement by 40%.
- Reduced core API latency by 60% with Redis caching and optimized data access, improving end to end response times.
- Owned a Kafka event pipeline handling more than 100K daily events with zero data loss, ensuring reliable analytics and business workflows.
- Enhanced worklog management UI in Angular with keyboard navigation and ARIA labels, achieving WCAG 2.1 compliance.
- Built Grafana and Prometheus dashboards and alerts, cutting incident response time by 50% and improving service reliability.
- Implemented Azure Blob Storage for document management, handling more than 50K contractor files with automated lifecycle policies.
JavaSpring BootKafkaRedisPostgreSQLAngularGrafanaPrometheusAzure Blob StorageDockerKubernetesAWSAzure - 2021-08 — 2021-12Tanu Trading & Co (Internship)Software Engineer InternAndhra Pradesh, India
- Developed inventory management APIs in C++, REST, and PostgreSQL handling more than 50K requests per day with p99 latency under 200 ms.
- Built an authentication layer using C++, JWT, and OpenSSL that enforced role based access control for more than 500 users.
- Deployed a serverless reporting pipeline with AWS Lambda, S3, and CloudWatch to process inventory snapshots with 99.9% reliability.
- Automated reporting jobs with SQL, PostgreSQL, and Cron, reducing stockouts by 25% and eliminating 12 hours per week of manual work.
C++RESTPostgreSQLJWTOpenSSLAWSLambdaS3CloudWatchSQLCron