Courses
From AI foundations to principal-level engineeringAI software architect - transition in 9 months
A 9-month, 39-week learning series.
Python-first · hands-on builds · architecture reasoning · no prerequisite vendor course
39
Weeks
6
Skill domains
4
Mastery stages
Coming soon
AI software architect - transition in 9 months
You're on the waitlist!
We'll email you the moment the course goes live.
Something went wrong. Please try again or email vinitk805582@gmail.com.
Six learning domains
What you will master over 39 weeks.
LLM & Model Engineering
Transformers, tokenization, embeddings, open-weight models, PEFT, inference behavior
RAG & Data Systems
Hybrid search, reranking, ingestion, data lineage, tenant security, multimodal retrieval
Agents & Harness Engineering
LangGraph, MCP, durable runtimes, context, memory, skills, coding agents, sandboxes
Eval, Security & Reliability
Golden datasets, trajectory evals, LLMOps, identity, policy, observability, recovery
AI Infrastructure & Platforms
vLLM, model gateways, GPU serving, advanced EKS, scaling, latency and cost
Principal Python & System Design
Typing, async, API contracts, LLD, databases, testing, technical strategy
Depth = target proficiency at the end of the series, not proficiency at the start.
The skills ladder
Milestone by milestone. Each one adds capability; earlier skills continue through later projects and design reviews.
-
W01–W08
Foundations
L1 · Build
Explain LLM behavior; write typed Python APIs; implement secure enterprise RAG.
-
W09–W18
Agent Engineer
L2 · Orchestrate
Design durable agents, MCP tools, memory, harnesses, evaluations and execution security.
-
W19–W27
Platform Engineer
L3 · Operate
Operate gateways and inference, optimize latency/cost, integrate EKS and harden the platform.
-
W28–W39
AI Software Architect
L4 · Lead
Defend system architectures; address scale, security, resilience, migrations and principal-level decisions.
How the learning time is allocated
18 hours of AI + 12 hours of Python and software engineering every week.
60%
AI architecture & engineering
LLMs · RAG · agents · harnesses · evaluations · inference · security
40%
Python & software engineering
APIs · concurrency · data modeling · system design · testing · performance
Week by week
Learn → implement → test → demonstrate mastery. Every week has a hands-on assessment and ends with one demonstrable deliverable.
W01–W08 · L1 · Build
Foundations
W01–W08 · L1 · Build
Foundations
AI architecture and requirements
What you will learn
Architecture patterns, quality/latency/cost tradeoffs, C4 and ADR
What you will build in Python
Python project scaffolding, strict typing, API contracts
End-of-week project
Architecture options + typed Python repo
LLM internals and token economics
What you will learn
Transformers, embeddings, sampling, tokenization, KV cache and TTFT
What you will build in Python
Async Python model client, task cancellation, exceptions
End-of-week project
LLM token, latency & cost profiler
Model APIs, structured outputs and model evaluation
What you will learn
Function calling, structured output, routing, streaming, provider capability matrix
What you will build in Python
Protocols, generics, Pydantic, error taxonomy and API interface design
End-of-week project
Typed multi-model comparison service
Open-weight models and fine-tuning choice
What you will learn
Transformers inference, quantization, LoRA/PEFT, quality-vs-cost decisions
What you will build in Python
Packaging and CLI design; Python profiling baseline
End-of-week project
Open-model benchmark + RAG/LoRA ADR
Vector retrieval and ingestion
What you will learn
Embeddings, chunking, HNSW, pgvector/Qdrant, metadata
What you will build in Python
SQLAlchemy async transactions, typed ingestion service
End-of-week project
Tenant-aware document search service
Hybrid retrieval and reranking
What you will learn
BM25, dense+sparse search, reranking, contextual retrieval
What you will build in Python
Bounded concurrency, rate limiting and profiling
End-of-week project
Hybrid search + reranker benchmark
Enterprise knowledge controls
What you will learn
ACL security, tenant isolation, index refresh, provenance
What you will build in Python
Repository authorization, secure multi-tenant transactions
End-of-week project
Permission-aware incremental RAG
RAG evaluation and multimodal introduction
What you will learn
Faithfulness, groundedness, GraphRAG and multimodal document extraction
What you will build in Python
pytest/Hypothesis evaluation fixtures, labeled datasets
End-of-week project
RAG evaluation + multimodal ingestion
W09–W18 · L2 · Orchestrate
Agent Engineer
W09–W18 · L2 · Orchestrate
Agent Engineer
Agent patterns and simple execution
What you will learn
Workflow-vs-agent, routing, tools, reflection, deterministic alternatives
What you will build in Python
SOLID, strategy, ports and adapters in Python
End-of-week project
Workflow vs agent implementation test
Durable agent runtime basics
What you will learn
LangGraph persistence, checkpoints, replay, durable workflows
What you will build in Python
Idempotency, transactional outbox, state machines
End-of-week project
Crash-resumable agent task runner
Context engineering and advanced memory
What you will learn
Context assembly, compaction, budgeting, provenance, write/read memory
What you will build in Python
Data models, domain contracts, context manager interface
End-of-week project
Context compression strategy benchmark
Multi-agent planning and MCP
What you will learn
Supervisor-worker, handoffs, secure tools, MCP auth and approvals
What you will build in Python
Secure Python service and tool contracts
End-of-week project
Multi-agent pipeline + secure MCP tools
Agent skills and protocol interoperability
What you will learn
Skills, progressive disclosure, agent-tool schemas, A2A interoperability
What you will build in Python
Python interfaces, graph data structure and design patterns
End-of-week project
Reusable agent skills + tool contracts
Harness foundations and repository engineering
What you will learn
Harness architecture, repo legibility, AGENTS.md, tool policies, verification
What you will build in Python
Architectural invariants enforced by Ruff/Pyright/pytest
End-of-week project
Coding-agent repo harness + policy checks
Coding agent SDLC and long-running tasks
What you will learn
Agent-driven coding, PR review, worktrees, repair loops and session recovery
What you will build in Python
Code review standards, artifact provenance, CI policy
End-of-week project
Autonomous patch-test-review loop
Advanced agent evaluations and statistical rigor
What you will learn
Trajectory correctness, task environment fixtures, judge calibration, confidence intervals
What you will build in Python
pytest property tests and repeatable eval harness
End-of-week project
Agent task evaluation and judge calibration
AI security, sandboxing and identity
What you will learn
Injection, tool exfiltration, least privilege, sandboxing, delegated identity
What you will build in Python
Secure Python IO, object authorization, secret handling
End-of-week project
Sandboxed tool execution + security tests
AI data provenance, supply chain and governance
What you will learn
Dataset/version lineage, model/tool provenance, supply-chain integrity, NIST
What you will build in Python
Python reproducible builds and typed dataset schemas
End-of-week project
Dataset provenance + software supply-chain report
W19–W27 · L3 · Operate
Platform Engineer
W19–W27 · L3 · Operate
Platform Engineer
Agent observability and AI reliability
What you will learn
Semantic tracing, outcome SLOs, cost, queue behavior, incident response
What you will build in Python
Context propagation, profiling, distributed Python workers
End-of-week project
End-to-end AI traces + tenant cost dashboard
Inference architecture, serving and optimization
What you will learn
vLLM, continuous batching, KV cache, GPU economics, provider selection
What you will build in Python
Python benchmarking, memory/profiling and concurrent clients
End-of-week project
Inference throughput and GPU/EKS design
AI gateway and enterprise platform design
What you will learn
LiteLLM routing, quotas, caching, security, failover, AI service catalog
What you will build in Python
FastAPI gateway design, versioned contracts, circuit breakers
End-of-week project
Multi-provider gateway + capacity model
Structured-data agents and multimodal capabilities
What you will learn
Text-to-SQL, safe execution, schema grounding, document+vision inputs
What you will build in Python
SQLAlchemy readonly boundaries, typed plugins, package design
End-of-week project
Secure text-to-SQL + multimodal route
Integrated enterprise platform build
What you will learn
Connect retrieval, gateway, agent runtime, harness, security and eval layers
What you will build in Python
Python modular architecture, OpenAPI, SDK and versioned contracts
End-of-week project
Integrated agent platform + Python SDK
Resilience and platform hardening
What you will learn
Tenant isolation, SLOs, failure recovery, cost attribution, release gates
What you will build in Python
Hyperscaler AI platform contracts (Google Cloud) and Python integration patterns
End-of-week project
Load, recovery & adversarial test suite
Principal-level system design and interviews
What you will learn
Build-vs-buy, migration design, scaling and AI governance decisions
What you will build in Python
Python DSA, LLD, design interviews, code reviews
End-of-week project
AI architecture case studies + coding tasks
Leadership, portfolio and roadmap defence
What you will learn
Architecture strategy, risk committee review, ROI, continuous feedback loops
What you will build in Python
Hyperscaler AI platform contracts (Microsoft Foundry, Azure) and Python integration patterns
End-of-week project
Principal engineering RFC + platform demo
Inference integration & Python performance
What you will learn
Serving final benchmark / memory-profile capstone inference
What you will build in Python
Python concurrency, py-spy, memory profiling and worker cancellations
End-of-week project
Managed vs self-hosted inference benchmark
W28–W39 · L4 · Lead
AI Software Architect
W28–W39 · L4 · Lead
AI Software Architect
Product requirements and architecture framing
What you will learn
Product requirements, AI success metrics, design alternatives
What you will build in Python
Python domain model, RFC quality, maintainable interfaces
End-of-week project
Enterprise AI product RFC and SLOs
Capacity and API design
What you will learn
Capacity planning, scale-out design, API boundaries
What you will build in Python
Async Python API contracts, concurrency and safe pagination
End-of-week project
10x capacity model + versioned Python API
Data architecture and enterprise retrieval
What you will learn
Search+OLTP+vector database design and data lifecycle
What you will build in Python
Python ORM optimization, transactions, migrations
End-of-week project
Data isolation + evaluated text-to-SQL
Messaging and workflow correctness
What you will learn
Event ordering, eventual consistency, outbox and saga
What you will build in Python
Python worker scheduling, idempotency, transactional outbox
End-of-week project
Event-driven job processor & replay test
Reliability at multi-tenant scale
What you will learn
SLOs, chaos testing, tenant noise isolation, degradation
What you will build in Python
Python retry budgets, rate limiting, circuit breakers
End-of-week project
Multi-tenant chaos and failure injection
Threat modeling and trusted control plane
What you will learn
Delegated identity, MCP authorization, secrets handling
What you will build in Python
Secure Python SDK and authenticated service integration
End-of-week project
Identity threat model + artifact provenance
Real-time agent events and observability
What you will learn
Streaming, browser notifications, audit events
What you will build in Python
Python SSE/WebSocket, queue consistency, contract testing
End-of-week project
Streaming agent events + reconnect replay
Search/recommendation design interviews
What you will learn
Search architecture, ranking and vector/hybrid retrieval scaling
What you will build in Python
Python ranking plugins, retrieval eval harness
End-of-week project
Relevance improvements + system-design cases
Collaboration and multimedia product systems
What you will learn
Document sync, multimodal data, content collaboration
What you will build in Python
Typed Python ingestion modules, queues and storage interfaces
End-of-week project
Multimodal ingestion & metadata versioning
Payments, auditability, change management
What you will learn
Idempotency, audit logs, transactional ownership
What you will build in Python
Python state machines, failure recovery, tests
End-of-week project
Human approval pipeline + tamper-evident audit
Migration and continual improvement strategy
What you will learn
Platform evolution, technology selection, ROI and feedback loops
What you will build in Python
Python packaging and API versioning; migration tests
End-of-week project
Platform migration roadmap + rollout experiment
Principal engineer capstone defense
What you will learn
Integrated AI runtime, architecture tradeoffs, scale, delivery strategy
What you will build in Python
Python review, DSA/LLD mock interviews, mentorship artifacts
End-of-week project
Final end-to-end platform defense and demo
Each build ends with an artifact, tests, evaluation evidence or an architecture review.
Who is this series for?
Open to professionals across the software industry. A shared learning path — from existing software-industry skills to AI engineering and architecture.
Software developers
Build AI-native products, APIs and applications.
Cloud & platform engineers
Extend cloud expertise into enterprise AI platforms.
DevOps engineers
Apply delivery and automation expertise to agentic systems.
Site reliability engineers
Engineer dependable, observable and scalable AI services.
QA & test automation
Develop evaluations, quality gates and agent test harnesses.
Data, security & other IT roles
Contribute data, governance and risk expertise to AI systems.
No single professional background is required; learners can deepen their existing strengths.
Join the waitlist