a
arocksharon32

Arock

@arocksharon32

Software Developer

Indien
Tamil, Englisch
Einige Informationen werden in englischer Sprache angezeigt.
Über mich
Hi, I am a passionate writer and developer who thrives on crafting engaging content and building innovative digital solutions. I have written and published 8 articles on LinkedIn, delivering insightful and well researched pieces. On the tech side, I have developed a ChatGPT-like app using the OpenAI API, showcasing my expertise in AI powered applications. Whether you need compelling articles, a sleek website, or a custom-built app, I’m here to turn your ideas into reality with quality and creativity. Let’s collaborate.... Mehr lesen

Kompetenzen

a
arocksharon32
Arock
offline • 
Durchschnittliche Antwortzeit: 1 Stunde

Meine Dienstleistungen

Professionelle Fachartikel
I will write high quality articles and blog posts that engage readers
Ab5 $ / ProjektWeitere Details
Entwicklung von KI-Chatbots
I will build a custom ai chatbot trained on your business documents
Ab10 $ / ProjektWeitere Details

Portfolio

Arbeitserfahrung

Amazon

Software Developer Intern

Amazon • Vollzeit

Jan 2026 - Jun 2026 • 5 mos

◦ Ported an iOS image-search matching algorithm to Android with full functional parity, implementing Lucene-based tokenization and proximity-gap matching and integrating a native C++ search library (KSDK BookSearch) via JNI, including patching the native build configuration to produce an Android-compatible binary. ◦ Built the Jetpack Compose UI layer for the ported search feature: a reactive carousel driven by StateFlow, asynchronous image loading via Coil, and full localization and accessibility support across 9 locales. ◦ Designed and built a semantic search proof-of-concept end-to-end on AWS: generated embeddings with Bedrock (Cohere Embed v3), indexed in OpenSearch Serverless, and configured two Bedrock Knowledge Bases (Titan TextEmbed v2) for auxiliary content types; implemented the retrieval Lambda, API Gateway integration, and client UI, unifying search across three previously separate content types. ◦ Shipped an entity-highlighting (”Hotspots”) feature surfacing contextual information for in-book entities, using a sidecar JSON schema for position and metadata.