Arpan Chatterjee

Software Development Engineer II at Amazon · Seattle, WA

arpanchatterjee2312@gmail.com · LinkedIn · Portfolio · Download PDF

Profile

Software engineer at Amazon with 7+ years across production AI, ML platforms, and distributed systems. Owned engineering for a contract-intelligence application used by 50+ legal professionals; built self-service ML evaluation and led multi-tenant ad-serving architecture across 5–6 engineers. Combines backend depth, full-stack delivery, and product judgment from stakeholder discovery through operations.

Experience

Amazon — Prime Video

Software Development Engineer II · Apr 2025 — Present · Seattle, WA

Owned model deployment, backend, and frontend for an AI-powered contract-intelligence application used by 50+ legal professionals at Amazon; productionized infrastructure, access controls, and monitoring.
Designed and built self-service ML evaluation: 10 production models and 80+ experimental versions, comparing accuracy, precision, recall, and F1 without platform-team intervention.
Own end-to-end development of an audio-description planning and vendor-allocation platform: stakeholder requirements, UX, architecture, frontend, backend, ordering, and fulfillment tracking.
Built localization recommendation pipelines for global studios and content partners; re-architected ~1.4 GB model-output processing, achieved a 99.91% validation pass rate, and integrated asset-generation ordering.
Built a streaming UI-generation backend on Amazon Bedrock AgentCore; complex prompts completed in 67 seconds after previously timing out at 300+ seconds.

Amazon — Amazon Ads

Software Development Engineer II · Mar 2023 — Apr 2025 · Seattle, WA

Architected and led Shared Cell implementation across 5–6 engineers. Established the standard multi-tenant production model, including isolation for throttling, logging, observability, and data access.
Led live-event performance engineering at 15M+ concurrent-viewer scale. Improved a worst-case path 3× (75→225 TPS/host), reducing required capacity ~67%; broader work helped avoid an estimated $10–15M in seasonal TNF EC2 costs.
Designed the AWS Graviton migration: 250→500 TPS/host, ~75% lower cost/request, and ~$1.4M projected annual EC2 savings without degrading latency.
Built reusable backend capabilities for Prime Video, Freevee, Twitch, IMDb, Goodreads, Alexa, and live-sports launches, contributing to $100M+ in revenue impact.

Amazon — Transportation Services

Software Development Engineer II · Aug 2022 — Mar 2023 · Seattle, WA

Built a facility-operations visibility platform for warehouse managers to track package state and movement, identify shipments at risk of missing deadlines, and prioritize operational interventions.

Amazon — Prime Video India

Software Development Engineer II · Oct 2021 — Aug 2022 · Bengaluru, India

Implemented Channels subscription e-mandates for RBI-required recurring-payment compliance and business continuity.
Built Free Trial Eligibility Service capabilities to detect repeat free-trial abuse using duplicate payment methods and other customer signals.

Amazon — Prime Video India

Software Development Engineer I · Aug 2019 — Sep 2021 · Bengaluru, India

Built payment-data migration and localization workflows supporting RBI data-residency requirements and recurring-payment e-mandates for Prime subscriptions in India.
Contributed to the India Channels launch, including third-party onboarding such as Eros Now and Discovery+.
Added Simplified and Traditional Chinese keyboard functionality to the Prime Video Living Room app.

National Instruments — Research & development

R&D Intern · Jan 2019 — Jun 2019 · Bengaluru, India

Developed automation to generate electronic circuit designs from user-specified inputs, outputs, and computation requirements.

Capabilities

Software & systems: Backend systems, full-stack applications, distributed architecture, multi-tenancy, performance profiling, capacity engineering, AWS.

ML & applied AI: Evaluation platforms, model-output pipelines, data-quality validation, AI-agent productionization, Amazon Bedrock AgentCore.

Technical delivery: Stakeholder discovery, UX and workflow design, architecture, cross-functional partnership, implementation, testing, deployment, onboarding, production operations.

AI-native engineering

Prototyping agent-driven software delivery from requirements and design through implementation and validation. Built independent validator-agent checks, blocking pre-review gates, and repository-specific rules to evaluate reduced human intervention with automated quality controls.

Education

Bachelor of Engineering in Computer Science
Bangalore Institute of Technology · 2015–2019

Languages: English, Hindi, Bengali