12-Week Learning Plan

AI/Backend Career Learning Plan

Target Roles (Priority Order):

1. AI Application Engineer / GenAI Engineer
2. AI/ML Engineer
3. Python/FastAPI Backend Engineer
4. Solutions Engineer

Timeline: 12 Weeks Approach: Project-Based Learning + DSA Practice

---

Role Priority & Overlap Analysis

                    AI/ML Foundations    Python/FastAPI    Cloud/Deployment    Communication
─────────────────────────────────────────────────────────────────────────────────────────────
AI App Engineer         ████████████        ████████████       ██████████         ████████
AI/ML Engineer          ████████████████    ██████████         ████████           ██████
Python Backend          ████                ████████████████   ████████████       ████
Solutions Engineer      ██████              ████████████       ██████████         ████████████████

Key Insight: ~60% of skills overlap across all four roles. Learn once, apply everywhere.

---

DevOps/Cloud Relevance

Role	Relevance	What You Need
AI App Engineer	⭐⭐⭐⭐ High	Docker, AWS (SageMaker, Lambda, Bedrock), basic K8s
AI/ML Engineer	⭐⭐⭐⭐⭐ Very High	MLOps, model deployment, GPU instances, K8s for ML
Python Backend	⭐⭐⭐ Medium	Docker, basic AWS (EC2, RDS, S3), CI/CD
Solutions Engineer	⭐⭐ Low-Medium	Conceptual understanding, not hands-on

Recommendation: Learn Docker + AWS basics (EC2, S3, Lambda) as foundation. Deep Kubernetes later if needed.

---

Daily Structure

┌─────────────────────────────────────────────────────────────────┐
│  Morning (1-2 hrs)    │  DSA: 1 Neetcode problem               │
│                       │  → Build habit, stay interview-ready   │
├───────────────────────┼─────────────────────────────────────────┤
│  Afternoon (2-3 hrs)  │  Learn: ONE concept deeply             │
│                       │  → Videos, docs, tutorials             │
├───────────────────────┼─────────────────────────────────────────┤
│  Evening (2-3 hrs)    │  Build: Apply to chatbot project       │
│                       │  → Hands-on implementation             │
└───────────────────────┴─────────────────────────────────────────┘

Weekly Total: 25-30 hours

---

12-Week Learning Plan

Phase 1: AI Core (Weeks 1-4)

Focus: Roles #1 and #2

Week 1-2: LLM Fundamentals

Day	DSA	Learn	Build
Mon	Arrays - Contains Duplicate	How LLMs work (transformers basics)	-
Tue	Arrays - Valid Anagram	Tokenization & embeddings	-
Wed	Arrays - Two Sum	Prompt engineering fundamentals	Test different prompts on chatbot
Thu	Arrays - Group Anagrams	Chain-of-thought, few-shot prompting	Improve chatbot system prompt
Fri	Arrays - Top K Frequent	LangChain intro - Chains	Start LangChain refactor
Sat	-	LangChain - Agents	Continue LangChain integration
Sun	-	Review & rest	Cleanup & commit

Resources:

• Andrej Karpathy’s “Let’s build GPT” (YouTube)
• Jay Alammar’s “The Illustrated Transformer”
• OpenAI Prompt Engineering Guide
• LangChain Documentation

Week 3-4: RAG & Vector Databases

Topic	Learn	Build
Chunking Strategies	Semantic vs fixed-size, overlap	Implement RecursiveCharacterTextSplitter
Embedding Models	OpenAI, Cohere, Sentence Transformers	Compare embedding quality
Retrieval Techniques	Similarity, MMR, hybrid search	Add BM25 + FAISS hybrid
Vector DBs	Pinecone vs Chroma vs Weaviate	Migrate to Chroma (or keep FAISS)
Evaluation	Retrieval accuracy metrics	Build eval pipeline
Citations	Confidence scores, source tracking	Add relevance scores to UI

Resources:

• LangChain RAG tutorials
• Pinecone Learning Center
• “Building RAG Applications” - DeepLearning.AI

---

Phase 2: Advanced AI + Backend (Weeks 5-8)

Focus: Roles #1, #2, and #3

Week 5-6: AI Agents & Tools

Topic	Learn	Build
ReAct Pattern	Agent reasoning loops	Simple ReAct agent
Function Calling	OpenAI/Anthropic function calling	Add tool use to chatbot
LangGraph	Multi-step agent workflows	Complex agent pipeline
MCP	Model Context Protocol	MCP server basics
Multi-Agent	Agent communication patterns	Two-agent system

Resources:

• LangGraph Documentation
• Anthropic MCP Documentation
• “Building AI Agents” - Harrison Chase tutorials

Week 7-8: FastAPI Production Patterns

Topic	Learn	Build
Async Python	async/await deep dive	Async endpoints
FastAPI Advanced	Dependency injection, middleware	Refactor chatbot backend
Authentication	OAuth2, JWT tokens	Add auth to API
Database	SQLAlchemy async, Alembic	Add conversation storage
Caching	Redis basics	Cache embeddings/responses
Testing	pytest, async testing	Test suite for API

Resources:

• FastAPI Official Documentation
• TestDriven.io FastAPI tutorials
• ArjanCodes YouTube (Python best practices)

---

Phase 3: Deployment & Polish (Weeks 9-12)

Focus: All 4 roles

Week 9-10: Cloud & Deployment

Topic	Learn	Build
Docker	Containers, Dockerfile, multi-stage	Dockerize chatbot
Docker Compose	Multi-container apps	Redis + API + UI setup
AWS EC2	Instance setup, security groups	Deploy chatbot
AWS S3	Object storage	Store embeddings/files
AWS Lambda	Serverless functions	Serverless endpoint (optional)
AWS Bedrock	Managed LLMs	Switch to Bedrock (optional)

Resources:

• Docker Official Tutorial
• AWS Skill Builder (free)
• FreeCodeCamp AWS Course

Week 11-12: Solutions Skills & Portfolio

Topic	Learn	Build
API Design	OpenAPI, versioning, documentation	Polish API docs
Technical Writing	Documentation best practices	README, blog post
Demo Skills	Presentation, storytelling	Record demo video
Troubleshooting	5 Whys, systematic debugging	Document debugging process
Portfolio	GitHub profile optimization	Clean all repos

Deliverables:

• 2-3 polished GitHub projects
• Updated LinkedIn with projects
• Blog post about RAG learnings
• Demo video of chatbot

---

DSA Track: Neetcode 150

Strategy: 1 problem/day, pattern-based approach

Schedule

Weeks	Pattern	Problems	Count
1-2	Arrays & Hashing	Contains Duplicate, Valid Anagram, Two Sum, Group Anagrams, Top K Frequent, etc.	9
3-4	Two Pointers + Sliding Window	Valid Palindrome, 3Sum, Container with Water, Best Time to Buy Stock, etc.	7
5-6	Stack + Binary Search	Valid Parentheses, Min Stack, Binary Search, Search Rotated Array, etc.	11
7-8	Linked List + Trees	Reverse LL, Merge Two LL, Invert Tree, Max Depth, Same Tree, etc.	18
9-10	Tries + Heap/Priority Queue	Implement Trie, Word Search, Kth Largest, Task Scheduler, etc.	9
11-12	Graphs + DP Basics	Number of Islands, Clone Graph, Climbing Stairs, House Robber, etc.	15

Tools:

• Neetcode.io - Organized problem list with video explanations
• LeetCode - Practice platform
• Language: Python (aligns with AI/Backend focus)

Rules:

1. Time-box: 30 mins max per problem
2. If stuck after 20 mins, look at hints
3. If stuck after 30 mins, study the solution
4. Always understand the pattern, not just the solution
5. Review failed problems on weekends

---

Chatbot Upgrade Roadmap

Current State

• FAISS vector search
• Sentence Transformers embeddings
• Groq API for LLM
• Streamlit UI
• Basic chunking

Level 1: Quick Wins

• Semantic chunking with overlap
• Chat memory (conversation history)
• Hybrid search (BM25 + vector)

Level 2: Production Features

• FastAPI backend (separate from UI)
• Evaluation pipeline with metrics
• Citation confidence scores
• Rate limiting & caching

Level 3: Advanced

• LangChain/LangGraph migration
• Multi-source RAG (PDF, web, etc.)
• Docker deployment
• AWS hosting

---

Resources Master List

LLM/AI

Resource	Type	Link
DeepLearning.AI Short Courses	Course	https://www.deeplearning.ai/short-courses/
LangChain Academy	Course	https://academy.langchain.com/
Hugging Face NLP Course	Course	https://huggingface.co/learn/nlp-course
Andrej Karpathy YouTube	Videos	https://www.youtube.com/@AndrejKarpathy
LangChain Docs	Docs	https://python.langchain.com/docs/

Python/FastAPI

Resource	Type	Link
FastAPI Documentation	Docs	https://fastapi.tiangolo.com/
TestDriven.io	Tutorials	https://testdriven.io/blog/topics/fastapi/
ArjanCodes	YouTube	https://www.youtube.com/@ArjanCodes
Real Python	Articles	https://realpython.com/

Cloud/DevOps

Resource	Type	Link
AWS Skill Builder	Course	https://skillbuilder.aws/
Docker Tutorial	Docs	https://docs.docker.com/get-started/
FreeCodeCamp AWS	YouTube	https://www.youtube.com/freecodecamp

DSA

Resource	Type	Link
Neetcode.io	Problems	https://neetcode.io/
LeetCode	Practice	https://leetcode.com/
NeetCode YouTube	Videos	https://www.youtube.com/@NeetCode

---

Weekly Checklist Template

## Week [X] - [Date Range]
 
### Goals
- [ ] Main learning goal
- [ ] Main building goal
- [ ] DSA problems (5-6)
 
### DSA Progress
- [ ] Mon: Problem name
- [ ] Tue: Problem name
- [ ] Wed: Problem name
- [ ] Thu: Problem name
- [ ] Fri: Problem name
 
### Learning Completed
- [ ] Topic 1
- [ ] Topic 2
 
### Built/Implemented
- [ ] Feature 1
- [ ] Feature 2
 
### Blockers/Notes
-
 
### Next Week Preview
-

---

Interview Talking Points

After completing this plan, you’ll be able to say:

“I built a RAG chatbot and iteratively improved it:

• Started with basic FAISS and chunking
• Added semantic chunking, improving retrieval accuracy by X%
• Implemented hybrid search for edge cases
• Built a FastAPI backend with auth and rate limiting
• Deployed on AWS with Docker
• Created an evaluation pipeline to measure quality”

This demonstrates: growth mindset, practical engineering, production thinking, and measurable impact.

---

Progress Tracker

Week	Phase	Status	Notes
1	LLM Fundamentals	⬜ Not Started
2	LLM Fundamentals	⬜ Not Started
3	RAG & Vector DBs	⬜ Not Started
4	RAG & Vector DBs	⬜ Not Started
5	AI Agents	⬜ Not Started
6	AI Agents	⬜ Not Started
7	FastAPI	⬜ Not Started
8	FastAPI	⬜ Not Started
9	Deployment	⬜ Not Started
10	Deployment	⬜ Not Started
11	Polish	⬜ Not Started
12	Polish	⬜ Not Started

---

Last Updated: February 2026 Created with Claude Code