Python · SQL · Linux · Git
Software engineering habits that make AI work maintainable.
MY JOURNEY / A CONNECTED UNIVERSE
Every step.
A new constellation.
Explore the milestones that connect my learning, work, products and research. Select a planet or choose a milestone below.
Every milestone is available below. The constellation is an optional way to explore.
Drag to orbit · scroll to zoom · select a planet2024–2028
Computer science foundations
B.Tech in Computer Science at Nagarjuna College of Engineering and Technology. Programming, databases, data structures, Linux and web development form the foundation.
Explore this chapter ↗THE COMPLETE JOURNEY / EXPERIENCE & MILESTONES
Journey / engineering path
From full-stack foundations to AI systems.
The through-line is practical: learn the minimum theory needed to build, debug, optimize and explain real systems.
Computer science + software foundations
Programming, databases, data structures, Linux, web development and security practice formed the base layer.
Applied ML + security automation
Projects moved into scikit-learn workflows, classification/forecasting, reconnaissance tooling and hands-on security labs.
GenAI systems
RAG, local LLMs, FastAPI, Ollama, prompt architecture and agent-oriented workflows became a central engineering focus.
Deep learning + production AI
Current work is moving through PyTorch and neural networks toward transformers, embeddings, fine-tuning, inference, evaluation and deployment.
Current stack
The engineering layers.
NumPy · Pandas · scikit-learn
Data preparation, modelling, evaluation and practical experimentation.
PyTorch · neural networks
Training loops, architectures, losses, optimizers and debugging.
Transformers · RAG · agents
LLM application architecture, retrieval, tool use and orchestration.
FastAPI · Docker · cloud
Serving, deployment, APIs and the infrastructure around models.
Nmap · Burp · Kali
Offensive-security practice used as a design and debugging lens.
Rule for the roadmap: if a topic does not improve my ability to build, debug, explain, optimize or deploy an AI system, it probably belongs later.
THE LEARNING PATH / ALL 28 PHASES
From foundations
to the flagship build.
00COMPLETED01COMPLETED02CURRENT03NEXT04UP NEXT05MUST KNOW06MUST KNOW07SUPPORTING08FOUNDATION09CORE10CORE11GENAI CORE12CORE13CORE JOB SKILL14TOOLS15CORE16CORE17ORCHESTRATION18ADVANCED19MODERN AI20DGX CORE21OPTIMIZATION22PRODUCTION23PRODUCTION24SUPPORTING25INTERVIEW CORE26MUST KNOW27FINAL BUILD
Phase 0 — Foundations
Python · Git/GitHub · SQL · Machine Learning
Phase 1 — Neural Network Foundations
Neuron → forward pass → loss → gradients → backpropagation
Phase 2 — Activation Functions
Step · Sigmoid · Tanh · ReLU · Leaky ReLU · Softmax
Phase 3 — Multi-Layer Neural Network
Build a complete network from scratch
Phase 4 — NumPy for AI
Vectors · matrices · batches · vectorization
Phase 5 — PyTorch
Tensors · autograd · models · datasets · training loops
Phase 6 — CPU + GPU Engineering
CUDA · memory · throughput · benchmarking
Phase 7 — Computer Vision Support
CNN fundamentals without turning CV into a second curriculum
Phase 8 — NLP Foundations
Text representation → embeddings → sequence concepts
Phase 9 — Transformers
Attention · self-attention · encoder/decoder · architecture
Phase 10 — Hugging Face
Transformers · tokenizers · pretrained models · inference
Phase 11 — Local LLM Engineering
LLM internals · generation · local inference · benchmarking
Phase 12 — Embeddings + Vector Search
Dense vectors · similarity · top-k · vector databases
Phase 13 — RAG
Ingestion → chunking → embeddings → retrieval → grounded generation
Phase 14 — LangChain + LlamaIndex
Frameworks for models, prompts, retrieval and orchestration
Phase 15 — Function Calling + Tools
Schemas · tool execution · structured arguments · failures
Phase 16 — AI Agents
Reason → tool → observe → reason → act
Phase 17 — LangGraph
Stateful graphs · routing · loops · human-in-the-loop
Phase 18 — Multi-Agent Systems
Supervisor · router · parallel agents · shared state
Phase 19 — MCP
Servers · clients · tools · resources · discovery
Phase 20 — LoRA / QLoRA / PEFT
Parameter-efficient fine-tuning on DGX Spark
Phase 21 — Quantization + Inference Optimization
Precision · memory · latency · throughput
Phase 22 — LLM Serving
FastAPI · streaming · batching · model servers
Phase 23 — Docker + MLOps / LLMOps
Containers · tracking · logging · monitoring · reproducibility
Phase 24 — Cloud
Compute · storage · IAM · containers · GPU deployment
Phase 25 — AI System Design
Architecture · scalability · tradeoffs · production decisions
Phase 26 — AI Evaluation
Measure retrieval, generation, agents and production behavior
Phase 27 — Flagship Production AI System
RAG + Agents + MCP + Fine-tuning + Serving + Evaluation