Siva Koneru

Full Stack Software Engineer (Contract)

GoodNotes • Freelance

Nov 2025 - Apr 2026 • 5 mos

• Implemented production AI-powered image workflows for GoodNotes Text Documents across the frontend and backend, enabling users to generate images from natural-language prompts, modify existing images using prompts, and ask questions about image content. • Resolved 100+ product bugs across web, backend, and iOS, including issues affecting newly introduced functionality; helped reduce crashes by 80% and clear all high-priority product-tagged bugs. • Designed and implemented an automated Jira-to-Hermes workflow that runs daily, classifies new tickets, and creates tasks for the appropriate Hermes agent, reducing manual bug triage and task preparation effort by 40%. • Developed and integrated Python/FastAPI backend services and TypeScript/React frontend functionality as part of AI feature development and product maintenance. • Collaborated with engineering, product, and design teams to investigate production issues, scope fixes, and deliver reliable releases.

Senior Software Engineer

Upland Software • Full-time

Sep 2021 - Oct 2025 • 4 yrs 1 mo

• Served as the primary engineer and architect for major components of Panviva's enterprise knowledge platform, supporting 11,000+ registered users; architectural decisions were reviewed with the system architect and team lead. • Designed and helped implement the modernization of a legacy ASP.NET Blazor monolith into modular .NET and Python/FastAPI services with a React frontend, improving the platform's scalability and maintainability. • Designed and implemented the production RAG architecture using Azure OpenAI, Azure AI Search, semantic chunking, embeddings, and Cosmos DB, with retrieval validation to improve answer reliability. • Built production AI orchestration workflows for multi-step user queries, including prompt enhancement, answer generation, hallucination detection, summarization, and title generation. • Engineered prompt caching and prompt optimization strategies that reduced AI infrastructure costs by 35%, while also reducing average AI response time by 20%. • Built an automated evaluation framework using golden datasets and an LLM-as-judge approach to assess answer quality and detect hallucinations, using the evaluation as a release gate with product and QA teams. • Built New Relic dashboards and alerts for the RAG platform and improved operational visibility into AI performance and failures. • Trained five new engineers joining the India team, helping them become productive across the platform and its architecture.

Senior Software Engineer

Bloomberg • Full-time

Jun 2020 - Jun 2021 • 1 yr

• Designed and implemented distributed fixed-income analytics for Bloomberg's Trade Order Management Systems, processing high-volume FX trade data and powering analytics used directly within Bloomberg Terminal workflows. • Re-architected analytics calculations from SQL Server aggregations to Apache Druid, significantly improving the performance of metrics such as G-spread and Z-spread and reducing analytics processing latency by 60%. • Designed the end-to-end real-time data pipeline for ingesting incoming trades, validating data, and calculating analytics for trading workflows. • Built both backend services and React-based Launchpad widgets used by traders to interact with real-time fixed-income analytics. • Improved system observability and production troubleshooting through centralized monitoring and alerting with Splunk and Grafana. • Decommissioned 30+ legacy services by consolidating duplicated functionality and migrating workloads to newer technology, reducing the maintenance burden of the platform.