Software Development Engineer @ Amazon Alexa | Agentic AI Frameworks | LangChain, LangGraph, Bedrock | Prompt Engineering & LLM Automation
As a Software Development Engineer at Amazon, I lead the design and execution of scalable test automation systems, data pipelines, and LLM-driven evaluation frameworks for Alexa's audio experiences. My work has delivered measurable business impact—boosting feature accuracy from 4% to 94.78%, reducing latency by 650ms, and scaling test coverage by 500%—while upholding the highest engineering standards across privacy, performance, and personalization.
With over 7+ years of experience as an SDE across conversational AI, cloud-native development, and enterprise systems, I bring a rare intersection of technical depth, operational rigor, and cross-functional leadership.
🚀 Key Achievements at Amazon:
- Designed intelligent tools and automated pipelines to streamline Conversation To Play online Tests evaluations, cutting turnaround time from a week to under an hour
- Led test architecture for multi-turn features across teams, authored extensible frameworks, and mentored new engineers and QA on large-scale evaluation practices
- Improved data ingestion pipelines between Amazon Music and P13N through checkpointing, deduplication, and recovery mechanisms
|
|
|
|
|
|
|
|
|
|
|
|
|
|
- Led accuracy and latency optimization for multi-turn conversational AI features, increasing recommendation-to-play accuracy from ~4% to 94.78% and reducing median latency by ~650ms
- Designed scalable test automation pipelines using agentic AI (Bedrock), cutting test evaluation turnaround from 3–4 days to under 1 hour and expanding test coverage by over 500%
- Architected large-scale data ingestion workflows, improving reliability by eliminating payload limit failures and reducing operational overhead through automated checkpointing and recovery
- Maintained 3,000+ automated test cases for online Tests evaluation system, enabling early edge case detection and robust regression prevention
- Mentored new engineers and QA members, accelerating onboarding timelines and enabling confident independent contributions
- Engineered Digital Library Management System for African Libraries using AWS RDS, enhancing transactional efficiency and inventory management
- Led database transition from NoSQL Firebase to relational structure, streamlining management of <1GB datasets while achieving $12.41 monthly cost savings
- Developed dynamic web reports using Next.js for comprehensive library operations insights including inventory tracking and user engagement metrics
- Led training for 40+ consultants on data structures, algorithms, HANA certification, and OData/REST API
- Managed production issues and mentored team members, delivering timely solutions using OOP and SQL expertise
- Led team of 17 technical consultants, delivering multiple successful projects and receiving client appreciation
- Resolved critical production bugs under rigorous deadlines, providing optimal solutions for complex SAP systems
- Implemented end-to-end SDLC for multiple implementations, leading 7-person team through requirements gathering to post-go-live support
- Achieved 12% increase in client revenue by implementing Supplier Lifecycle Management system with customized MVC design pattern
- Implemented data exchange systems from Ariba to ERP using SOAP protocol, RFC, and AIF for error handling
Course Connect - Fault-Tolerant Distributed System
- Skills: Python, Django, AWS RDS Postgres, REST API, Leader Election
- Description: Developed fault-tolerant distributed full-stack web app on AWS EC2 with separate front-end, message broker, and compute nodes. Implemented concurrency, failure handling, and leader election showcasing advanced system design skills.
Movie Personality Quiz - AI-Powered Recommendations
- Skills: Next.js, Vercel, Cosine Similarity, TMDB API
- Description: Built full-stack web app with responsive design, utilizing cosine similarity algorithms and TMDB API for personalized movie recommendations based on personality traits.
Task Tracker - Full-Stack Productivity App
- Skills: Next.js, Vercel, Firebase, Authentication
- Description: Deployed powerful task management app with CRUD operations, advanced sorting, user authentication, and real-time Firebase integration.
Big Data Analysis Platform
- Skills: Apache Spark, Hadoop, HDFS, Kafka, PySpark
- Description: Built scalable big data platform using Spark and Hadoop for top-k calculations, analyzed logs with Kafka streams and PySpark for comprehensive exploratory data analysis.
Portfolio Allocation Analyzer
- Skills: React, Chart.js, Financial APIs, Data Visualization
- Description: Developed React web app enabling users to analyze stock asset allocation, track investment performance with interactive data visualizations and real-time market data.
Data Mining & Pattern Recognition
- Skills: Python, Pandas, PyCaret, Matplotlib, Seaborn, ML
- Description: Demonstrated advanced data mining and pattern recognition skills with Kaggle datasets, implementing clustering algorithms and predictive models for hotel classification.
Multithreaded Web Crawlers
- Skills: Python, Multithreading, Concurrency, Algorithm Design
- Description: Programmed advanced task scheduling algorithms and analyzed locking techniques in multithreaded web crawlers, optimizing performance and resource utilization.
- 🤖 Agentic AI Frameworks: Expert in LangChain, LangGraph, LangSmith for building intelligent agent workflows and multi-step reasoning systems
- 🧠 Prompt Engineering & LLM Optimization: Advanced prompt design, chain-of-thought reasoning, and LLM performance tuning across OpenAI, Claude, and Bedrock models
- ⚡ AI-Powered Test Automation: Building scalable evaluation pipelines using AWS Bedrock and agentic AI for conversational AI testing at enterprise scale
- 🔗 Multi-Agent Systems: Designing complex agent orchestration, tool integration, and workflow automation using LangFlow and modern AI frameworks
- 📊 LLM Evaluation & Monitoring: Creating comprehensive evaluation frameworks, A/B testing for AI systems, and production monitoring for conversational AI
- ☁️ Cloud-Native Development: AWS-first approach with containerization and microservices and AWS Bedrock
- 📊 Data Engineering: Large-scale data processing, pipeline optimization, and real-time streaming
- 🔧 System Design: Distributed systems, fault tolerance, and high-availability architectures
- 👥 Technical Leadership: Mentoring engineers, cross-team collaboration, and driving technical excellence
💡 Passionate about architecting scalable solutions that make a real impact. Always open to discussing innovative tech challenges!
📫 Let's Connect: dm.divya.mahajan@gmail.com