Self-Assessment
Internship Overview
From September 8 to December 9, 2024, I interned at Amazon Web Services (AWS) and immersed myself in the daily rhythm of a production-grade cloud team. The experience forced me to connect classroom theories with the realities of deploying resilient, secure, and cost-aware workloads.
My primary project was the Bandup IELTS Learning Platform, an AI-assisted learning experience that evaluates Writing and Speaking tasks and generates contextual flashcards. I owned several backend and AI integration streams that turned the concept into a working system students could interact with.
Key Achievements
Working on Bandup allowed me to sharpen both technical depth and product thinking:
Cloud Architecture & AWS Services
- Sketched and iterated on a serverless blueprint using AWS Lambda, API Gateway, and SQS to decouple workloads
- Hardened networking by carving VPC public/private subnets, NAT Gateways, and Security Groups aligned with AWS Well-Architected guidance
- Provisioned container workloads on Amazon ECS with Fargate to host auxiliary services
- Chose the right data stores—Amazon S3 for documents, DynamoDB for high-scale metadata, and ElastiCache (Redis) for caching test artifacts
- Experimented with Amazon Bedrock (Titan Embeddings) alongside Google Gemini API to bring AI insight closer to end users
Development & DevOps
- Implemented Python Lambda functions that grade submissions and trigger flashcard creation
- Built the RAG (Retrieval-Augmented Generation) flow that indexes materials and surfaces contextually relevant snippets
- Wired CI/CD pipelines through GitLab runners and AWS CodePipeline so every change moved predictably from commit to deployment
- Practiced Infrastructure as Code, keeping environments reproducible and reviewable
AI/ML Integration
- Used Gemini’s native audio pipeline to process speaking samples at roughly 72% lower cost than a custom stack
- Embedded lesson chunks with Amazon Titan Text Embeddings V2 to support semantic search and scoring explanations
- Iterated on prompt engineering tactics to align AI scoring outputs with IELTS descriptors
Work Ethics
I held myself accountable for every deliverable by:
- Closing tasks with production-ready quality instead of proof-of-concept shortcuts
- Respecting AWS security guardrails and consistently reviewing cost impact before rolling changes out
- Scheduling proactive syncs with mentors so blockers surfaced early and were resolved quickly
- Capturing architecture decisions, runbooks, and test evidence to reduce knowledge gaps for anyone inheriting my work
Self-Evaluation Criteria
To gauge how much I truly grew, I rated myself on metrics that mattered to the project and team:
| No. | Criteria | Description | Good | Fair | Average |
|---|
| 1 | Professional knowledge & skills | Applying AWS design patterns, selecting the right services, and delivering production-grade code | ✅ | ☐ | ☐ |
| 2 | Ability to learn | Picking up new AI services and DevOps tooling quickly enough to unlock project milestones | ✅ | ☐ | ☐ |
| 3 | Proactiveness | Investigating design options and AWS docs before escalating for help | ✅ | ☐ | ☐ |
| 4 | Sense of responsibility | Owning Lambda features end-to-end, from architecture discussions to deployment checklists | ✅ | ☐ | ☐ |
| 5 | Discipline | Staying aligned with sprint ceremonies, coding standards, and daily stand-ups | ☐ | ✅ | ☐ |
| 6 | Growth mindset | Treating code reviews and architectural feedback as chances to iterate fast | ✅ | ☐ | ☐ |
| 7 | Communication | Explaining trade-offs in docs and demos, while still improving presentation clarity | ☐ | ✅ | ☐ |
| 8 | Teamwork | Supporting mentors and peers, sharing test data, and pairing on tricky debugging sessions | ✅ | ☐ | ☐ |
| 9 | Professional conduct | Following AWS security expectations, respecting data privacy, and keeping commitments | ✅ | ☐ | ☐ |
| 10 | Problem-solving skills | Unblocking Lambda bugs, tracing performance issues, and driving the 72% audio-processing savings | ✅ | ☐ | ☐ |
| 11 | Contribution to project/team | Shipping four Lambda services plus documentation, demos, and a repeatable AI pipeline | ✅ | ☐ | ☐ |
| 12 | Overall | Holistic view of my readiness to contribute as a junior cloud engineer | ✅ | ☐ | ☐ |
Key Learnings
Technical Skills Gained
- Strengthened my mental model for building serverless-first systems on AWS
- Proved out AI/ML integration patterns that translate research ideas into usable product features
- Became confident writing structured, testable Python for Lambda runtimes
- Internalized how RAG pipelines, vector stores, and embeddings interact to serve real queries
Soft Skills Developed
- Crafted documentation that balances architectural detail with actionable steps
- Broke down ambiguous issues into root causes and tracked fixes methodically
- Evaluated every feature for cost impact, looking for savings before launch
- Adapted to enterprise collaboration rhythms, tools, and review processes
Areas for Improvement
- Discipline: Build firmer routines so even on hectic days I never miss status updates or deadlines
- Communication: Practice storytelling for technical demos to make non-technical listeners comfortable
- Problem-solving: Adopt lighter-weight runbooks for debugging so future incidents resolve faster
- Networking: Invest time connecting with more AWS teams to better understand adjacent services and opportunities
Gratitude
Thank you to the mentors who trusted me with real production responsibilities, the operations team that kept the environment stable while I iterated, and AWS for opening its doors to curious students like me. The lessons from this internship now shape how I plan, build, and communicate every new idea.