Harshita Kalani

Delivered a production-ready Kotlin Multiplatform (KMP) module by onboarding the Player SDK's Playback Info Resolver, enabling seamless Kotlin-to-Swift/JS/C++ transpilation and reducing platform-specific duplication. Built and deployed a Google Notebook–style RAG system on GCP that ingests documents, links, and Git repositories, enabling semantic search and LLM-based QA workflows using Ollama and vector retrieval. Developed and productionized an anomaly detection pipeline for playback telemetry, leveraging statistical modeling and ML techniques to detect QoE deviations and eliminate false positives across millions of events.

SDE I - Warner Bros. Discovery

(2024-07)

• Delivered a production-ready Kotlin Multiplatform (KMP) module by onboarding the Player SDK's Playback Info Resolver, enabling seamless Kotlin-to-Swift/JS/C++ transpilation and reducing platform-specific duplication.
• Built and deployed a Google Notebook–style RAG system on GCP that ingests documents, links, and Git repositories, enabling semantic search and LLM-based QA workflows using Ollama and vector retrieval.
• Developed and productionized an anomaly detection pipeline for playback telemetry, leveraging statistical modeling and ML techniques to detect QoE deviations and eliminate false positives across millions of events.

AI Data Annotation Specialist - Leading AI Company (NDA)

(2024-07)

• Performed large-scale data labeling and code annotation to curate high-quality datasets for AI/ML models.
• Evaluated AI-generated code for correctness, efficiency, edge cases, and best practices across multiple languages.
• Curated SWE benchmark datasets and annotated chatbot/IR outputs, supporting end-to-end LLM training and evaluation pipelines.

GSoC | Compiler Developer - LFortran | NumFOCUS

(2023-07)

• Contributing to LFortran, a modern interactive LLVM-based Fortran compiler with an AST → ASR → LLVM IR pipeline.
• Improved LFortran's compatibility by enabling successful compilation of many Fortran packages.
• Implemented core intrinsic functions, significantly expanding compiler completeness as a GSoC'24 contributor.
• Mentored new contributors and represented LFortran at the GSoC mentor summit'25 in Munich, Germany.

Bachelors of Technology - Computer Science and Engineering - Indian Institute of Technology Jodhpur (2020-11 - 2024-05)

Secondary & Higher Secondary Education - Maheshwari Girls Public School (2016-06 - 2020-04)