Applytic

AI-powered job application tracker that learns from your rejections.

🚀 Start Using Applytic

Live in production. Free to use.

Applytic tracks every job application you submit, detects patterns across rejections (which resume version converts best, which source channel works, which company sizes respond), and uses Amazon Bedrock to turn that data into actionable coaching - delivered as a chat interface and a weekly email digest.

Built end-to-end on AWS as a production-grade application. Every service is serverless, infrastructure is code, and every push auto-deploys via GitHub Actions.

---

Why I Built This

I was job hunting and had no data on why I was getting rejected. I had a spreadsheet with company names and "rejected" written next to most of them - but no signal on why. Was it my resume? The channel? The company size?

So I instrumented my own job search. Every application became a data point. After a few weeks I had enough data to see that my v1-generic resume had a 0% response rate from enterprise companies, while v3-ml-focused was getting interviews from startups via referrals. That's the kind of insight you can act on.

---

Features

Application tracking

• Log applications with company, role, source channel, resume version, company size, job description URL, follow-up date
• Kanban board with drag-and-drop status updates (Applied - Screened - Interview - Offer / Rejected)
• Click any card to view full detail, edit all fields, change status, see timeline
• Search by company/role, filter by source channel
• Color-coded left border per status on kanban cards for instant visual scanning
• Amber "Follow up" badge on overdue cards

Weekly goal tracking

• Set a weekly application target on the Dashboard
• Progress bar showing current week vs goal, turns green when met
• Streak counter - consecutive weeks hitting your goal (fire icon)
• Inline goal editing - click the pencil to update your target anytime

Follow-up reminders

• Attach a follow-up date to any application
• Daily email when applications are overdue for follow-up, grouped by user
• Never miss a follow-up again

Notes timeline

• Add timestamped notes to any application
• Full per-application audit trail
• Notes sorted oldest first for easy reading

CSV export and import

• Export all applications to a CSV with one click - no Lambda, pure client-side
• Import from CSV with full validation, preview, and error reporting
• Download import template to get started quickly

AI insight engine

• Pattern analysis across 6 dimensions: source channel, company size, resume version, role seniority, weekly velocity, status funnel
• Response rate computed per bucket - shows exactly which resume version or source is working
• AI coaching chat powered by Bedrock - answers questions like "why am I getting ghosted?" using your actual data as context
• Markdown rendering in chat responses
• Weekly email digest every Monday with stats + one AI-generated personalised tip

Resume version tracker

• Upload multiple PDF versions to S3 via presigned URLs
• Tag each application with which version was used
• Analytics shows conversion rate per version side-by-side

UI / UX

• Full dark mode with system preference detection, persisted to localStorage
• Mobile responsive - hamburger sidebar on small screens
• Loading skeletons on every page instead of blank states
• Meaningful empty states with calls to action
• Toast notifications (top-center)

---

Architecture

---

Key Engineering Decisions

Single-table DynamoDB - two entity types (APPLICATION, STATUS_EVENT) in one table with composite keys. Every query is O(1). No joins. Scales to millions of requests/sec with no configuration changes.

Pattern analysis before LLM - the insights Lambda computes structured metrics (response rates per bucket) before calling Bedrock. The LLM receives hard numbers as context, not raw records. This makes advice data-driven and specific rather than generic.

Serverless-first - no EC2, no containers, no idle servers. Lambda + API Gateway + DynamoDB means ~$0 at low volume and automatic scaling. EventBridge replaces cron servers entirely.

ARM64 Lambda - all functions run on Graviton2 (ARM64) for ~20% cost reduction and faster cold starts vs x86.

Presigned S3 URLs for resume upload - the Lambda generates a presigned URL and returns it to the client. The file then uploads directly from the browser to S3 - never passes through Lambda.

Lambda Layer for shared code - single source of truth for CORS, auth extraction, error formatting, Pydantic validation, and X-Ray tracing. All seven Lambda handlers share it.

Client-side CSV - export and import are handled entirely in the browser. No Lambda invocation needed for either operation, keeping them fast and free.

Streak computation - looks back 8 weeks maximum, current week excluded (may be incomplete). Walks back week by week, breaks on first miss.

---

Tech Stack

Layer	Service
Frontend	React 18, TypeScript, Vite, Tailwind CSS, react-markdown
Auth	Amazon Cognito (email + JWT)
API	API Gateway REST + Lambda (Python 3.12, ARM64)
AI / ML	Amazon Bedrock - Amazon Nova Lite
Database	DynamoDB - single-table design, PAY_PER_REQUEST
Storage	S3 - resume versioning + frontend hosting
CDN	CloudFront (450+ edge locations)
Scheduling	EventBridge cron (Monday 8am UTC digest, daily 9am UTC follow-ups)
Email	Amazon SES
IaC	AWS CDK v2 TypeScript
CI/CD	GitHub Actions

---

Project Structure

applytic/
├── cdk/                    # AWS CDK v2 stack - all infrastructure as code
│   ├── bin/app.ts
│   └── lib/applytic-stack.ts
├── lambdas/                # Python 3.12 Lambda handlers
│   ├── applications/       # CRUD + presigned S3 URL generation
│   ├── insights/           # Pattern analysis + Bedrock AI coaching
│   ├── digest/             # Weekly SES email digest
│   ├── followup/           # Daily follow-up reminder emails
│   ├── settings/           # Weekly goal + streak tracking
│   ├── notes/              # Per-application notes timeline
│   ├── cognito_verify/     # Post Confirmation trigger - SES email verification
│   └── shared_layer/       # Lambda Layer - shared middleware, Pydantic, X-Ray
├── frontend/               # React + Vite + Tailwind CSS
│   └── src/
│       ├── components/     # Kanban, Analytics, Chat, Sidebar, Resume, CSV
│       ├── hooks/          # useApplications, useSettings, useNotes
│       ├── lib/            # API client, Amplify config, theme, csv utils
│       ├── pages/          # Dashboard
│       └── types/          # Shared TypeScript types
├── tests/                  # 207 pytest tests (90.75% coverage)
│   ├── test_applications.py
│   ├── test_applications_integration.py
│   ├── test_insights.py
│   ├── test_insights_integration.py
│   ├── test_digest.py
│   ├── test_cognito_verify.py
│   ├── test_followup.py
│   ├── test_settings.py
│   └── test_notes.py
├── scripts/
│   ├── seed_data.py
│   ├── build_layer.sh
│   └── setup_oidc_v11.sh
└── .github/workflows/
    ├── deploy.yml
    └── codeql.yml

---

DynamoDB Single-Table Design

Table name: applytic

Entity	PK	SK	GSI1PK	GSI1SK
Application	USER#userId	APP#appId	USER#userId	DATE#timestamp
Status event	APP#appId	EVENT#timestamp	-	-
Rate limit	RATELIMIT#userId	DATE#YYYY-MM-DD	-	-
User settings	USER#userId	SETTINGS	-	-
Note	APP#appId	NOTE#timestamp#noteId	-	-

Access patterns:

• List all applications for user - GSI1 query on USER#userId sorted by date
• Get single application - Main table get on USER#userId + APP#appId
• Get status history for an app - Query APP#appId with EVENT# prefix
• Get notes for an app - Query APP#appId with NOTE# prefix
• Get user settings - Get on USER#userId + SETTINGS

---

API Routes

GET    /v1/applications
POST   /v1/applications
GET    /v1/applications/{appId}
PUT    /v1/applications/{appId}
DELETE /v1/applications/{appId}
POST   /v1/applications/{appId}/status
GET    /v1/applications/{appId}/notes
POST   /v1/applications/{appId}/notes
DELETE /v1/applications/{appId}/notes/{noteId}
POST   /v1/resumes/upload-url
GET    /v1/resumes/list
GET    /v1/insights
POST   /v1/insights/chat
GET    /v1/users/settings
PUT    /v1/users/settings

All routes protected by Cognito JWT authorizer.

---

Scale and Cost

Metric	Value
API throughput	10,000 req/sec (API Gateway default)
DynamoDB SLA	99.999% availability
Lambda cold start	~300-400ms (Python 3.12 ARM64)
CloudFront edge locations	450+ worldwide
Cost at 0 users	~$0/month
Cost at 100 users	~$2-5/month
Cost at 1,000 users	~$15-30/month
Backend test suite	207 tests, 90.75% coverage
Frontend test suite	14 Vitest tests
Full CDK deploy from scratch	under 3 minutes

---

CI/CD Pipeline

push to main
    ├── test (pytest 207 tests, 70% coverage threshold)
    ├── test-frontend (vitest 14 tests)
    └── (both must pass before)
            ├── deploy-backend (cdk deploy)
            ├── deploy-frontend-aws (s3 sync + cloudfront invalidation)
            └── deploy-frontend-pages (GitHub Pages)

---

Local Setup

Prerequisites

• Node.js 18+
• Python 3.12
• AWS CLI configured (aws configure)
• AWS CDK CLI: npm install -g aws-cdk

Deploy backend

bash scripts/build_layer.sh
cd cdk && npm install && cdk deploy

Run frontend locally

cd frontend && npm install && npm run dev

Run tests

pip install pytest boto3 pydantic aws-lambda-powertools pytest-cov
pip install "moto[dynamodb,s3,ses,cognitoidp]"
python -m pytest tests/ -v --tb=short --cov=lambdas --cov-fail-under=70

Run frontend tests

cd frontend && npm run test

Seed demo data

cd scripts && python seed_data.py --user-id YOUR_COGNITO_SUB

---

Issues Encountered and Fixed

1. TypeScript error on CDK deploy (Partial<FunctionProps> type)

• Issue: runtime became optional when using Partial<FunctionProps>, causing TypeScript to reject undefined values.
• Fix: Added runtime explicitly to each Lambda and removed the shared defaults spread.

2. esbuild crash on Windows

• Issue: esbuild failed due to corrupted dependencies or a Node.js version mismatch.
• Fix: Deleted node_modules and package-lock.json, reinstalled dependencies, and used Node 18/20 LTS.

3. pytest module name collision

• Issue: Multiple Lambda handlers were named handler.py, causing import collisions during tests.
• Fix: Used importlib.util.spec_from_file_location() with unique module names.

4. Timezone-aware datetime bug in velocity calculation

• Issue: dateApplied was stored as a YYYY-MM-DD string and could not be subtracted from a timezone-aware datetime.
• Fix: Detected 10-character date strings and added UTC timezone information before subtraction.

5. Chat history lost on navigation

• Issue: Component-level useState was reset whenever the component unmounted.
• Fix: Lifted message state to App.tsx and passed it down as props.

6. GitHub Actions npm cache error

• Issue: cache: 'npm' requires a package-lock.json file.
• Fix: Removed the cache configuration and used npm install.

7. import.meta.env TypeScript error in CI

• Issue: Vite client types were not available during CI builds.
• Fix: Added "types": ["vite/client"] to tsconfig.json.

8. Bedrock model EOL

• Issue: The selected Bedrock model reached AWS end-of-life during development.
• Fix: Updated BEDROCK_MODEL_ID and used aws logs tail for debugging.

9. Bedrock inference profile requirement

• Issue: Claude 3.7+ models require cross-region inference profiles using a us. prefix.
• Fix: Updated model configuration to use the correct inference profile identifier.

10. IAM wildcard required for inference profiles

• Issue: Foundation model ARNs did not cover inference profile ARNs.
• Fix: Used Resource: "*" for Bedrock IAM permissions.

11. AWS Marketplace subscription required

• Issue: Claude 3.7 and Haiku 4.5 required an AWS Marketplace subscription.
• Fix: Switched to Amazon Nova Lite.

12. Amazon Nova request/response format differences

• Issue: Nova models do not accept anthropic_version in the request payload.
• Fix: Detected the model family from MODEL_ID and generated the appropriate payload format.

13. Pie chart labels clipped

• Issue: External labels overflowed the chart container.
• Fix: Implemented a custom label renderer that draws percentage values inside slices at the midpoint radius.

14. Digest Lambda missing Cognito permission

• Issue: cognito-idp:AdminGetUser permission was missing from the IAM policy.
• Fix: Added an IAM policy statement targeting the Cognito User Pool ARN.

15. SES sandbox – recipient emails not verified

• Issue: New AWS accounts in the SES sandbox can only send emails to verified recipient addresses.
• Fix: Added a Cognito Post Confirmation trigger that automatically verifies newly registered user emails.

16. GitHub Pages blank page after login

• Issue: BrowserRouter defaulted to / while the application was hosted under /applytic/.
• Fix: Passed import.meta.env.BASE_URL as the basename to BrowserRouter.

17. Windows pip --user + -t conflict in build_layer.sh

• Issue: Windows automatically added --user, which conflicts with pip's -t option.
• Fix: Added the --no-user flag to the pip install command.

18. CDK construct ID mismatch on redeploy

• Issue: CloudFormation rejected deployment with a "Resource already exists" error.
• Fix: Preserved the original construct IDs when updating existing alarms.

19. Lambda ARM64 vs x86_64 architecture mismatch

• Issue: build_layer.sh produced x86_64 wheels while Lambdas ran on ARM64, causing pydantic_core runtime failures.
• Fix: Added --platform manylinux2014_aarch64 --only-binary=:all: to the pip install command.

20. parse_body stub missing error handling in test_insights.py

• Issue: The test stub crashed on invalid JSON input.
• Fix: Replaced the lambda with a proper function that wraps json.loads() in a try/except block and returns HTTP 400 on failure. The stub is registered in sys.modules and reused by subsequent test files.

21. AddApplicationModal missing followUpDate field

• Issue: TypeScript builds failed after v2.0 introduced followUpDate to the Application type.
• Fix: Added followUpDate: null as string | null to the defaultForm object.

22. moto installation failure in CI

• Issue: The shell interpreted brackets in moto[dynamodb,...] as glob expansion.
• Fix: Wrapped the package name in quotes during installation.

23. Vitest globals breaking tsc builds

• Issue: vi, beforeAll, and afterAll were not available during standard TypeScript compilation.
• Fix: Added "exclude": ["src/test"] to frontend/tsconfig.json.

---

Changelog

See CHANGELOG.md for full version history.

---

Contributing

See CONTRIBUTING.md for setup instructions, branching conventions, and the PR process.

---

Author

Hardik - Github