AI Engineering Training Programme
Six weeks, two parallel tracks. The Technical Track takes you from Python fundamentals to deploying a full-stack AI application. The Professional Skills Track builds the visibility, communication and career strategy that turn skill into opportunity. Every week ends with a shipped project, and a weekly live webinar brings in an industry guest on AI and integrations.
Jointly organised by Sephar-Innovations Ltd. & Innovation Eastern Action Company
6 weeks · Live cohort · 2 tracks/week · 6 hands-on projects · weekly industry webinar · deployed capstone
What you'll be able to do
- ✓ Build and deploy a full-stack AI application — your capstone, on a public URL
- ✓ Integrate LLMs, RAG, agents and tool-calling into real applications
- ✓ Run open-source models locally and fine-tune with LoRA / PEFT
- ✓ Communicate, position and present your work to earn opportunities
- ✓ Graduate with six shipped projects and a verified Certificate of Completion
Who it's for
- → Professionals entering AI engineering — no prior ML experience required
- → Developers adding production AI to their existing stack
- → Career-focused builders who want visibility, positioning and strategy
- → Teams upskilling together (corporate cohorts)
Tools & tech
6 weeks · 6 modules
Two parallel tracks every week — ⚙ Technical and ★ Professional Skills — each closing with a deliverable.
Python Foundations & API Integration
Week 1Building the core engineering toolkit for AI development.
- Python scripting essentials: data types, functions, file I/O, error handling
- Structuring and parsing JSON — the lingua franca of AI APIs
- REST API fundamentals: requests, responses, authentication
- Prompt engineering: zero-shot, few-shot, chain-of-thought
- Managing API keys, environment variables and rate limiting safely
Owning Your Professional Narrative
- Why most professionals undersell themselves — and how to stop
- A clear one-liner about what you do and who you serve
- LinkedIn profile audit: headline, summary and featured strategy
- Credentials vs positioning — and which moves faster
Deliverable: Rewritten LinkedIn headline + 3-sentence professional bio
Open-Source Models & Local Hosting
Week 2Running powerful AI without cloud dependency.
- The open-source AI landscape: model families, licences, use-case fit
- Model quantisation: what GGUF is and the trade-offs vs full precision
- Downloading, inspecting and validating models from Hugging Face Hub
- Hardware constraints: VRAM, RAM, CPU offloading and batch size
- Running local inference with Llama.cpp and Ollama
Building Visibility That Compounds
- How being known accelerates opportunity
- Learning in public — sharing work-in-progress without being an expert
- Platform selection: LinkedIn vs X vs newsletters vs GitHub
- Consistency over virality: a presence that grows with you
Deliverable: First 'learning in public' post drafted and scheduled
Retrieval-Augmented Generation (RAG)
Week 3Giving LLMs access to your data without retraining them.
- Why RAG exists: the problem with static knowledge cutoffs
- Text embeddings: how they encode semantic meaning
- Chunking strategies: fixed-size, semantic and hierarchical
- Vector stores: setting up ChromaDB and FAISS
- The pipeline: embed query → search → inject context → generate
- Evaluating retrieval: precision, recall and relevance
Communication That Earns Influence
- Translating technical work into business language
- The Pyramid Principle: lead with conclusions, not process
- Writing emails that get decisions, not questions
- Presenting to non-technical stakeholders without dumbing down
Deliverable: Rewrite a past technical update as an executive-ready brief
Agents & Tool Calling
Week 4Building AI systems that can reason, plan and act.
- The ReAct framework: thought → action → observation loops
- Function calling: defining tools an LLM can invoke at runtime
- Python functions with strict schemas for LLM-triggered execution
- Multi-agent patterns: orchestrator-worker, sequential, parallel
- Tool errors, retries and graceful degradation
- Safety and guardrails against runaway agent loops
Navigating Workplace Dynamics & Office Politics
- How organisations work beneath the org chart
- Reading a room: decision-makers, influencers and blockers
- Building alliances with a credibility-first approach
- Managing up; handling credit and recognition professionally
Deliverable: Stakeholder map with one underinvested relationship actioned
Fine-Tuning & Custom Workflows
Week 5Adapting models to your domain, style and data.
- When to fine-tune vs RAG vs prompt engineering
- Parameter-Efficient Fine-Tuning (PEFT): why it works
- LoRA in depth: rank decomposition, target layers, hyperparameters
- Dataset formatting: instruction-response pairs and data hygiene
- Running a fine-tuning job and monitoring loss curves
- The economics of training: when it makes business sense
Personal Brand Strategy & Positioning
- Reputation vs brand — and why both matter
- Finding your niche: AI skills × domain knowledge × personality
- A content strategy you can sustain for 12 months
- Monetising a personal brand: consulting, training, advisory, product
Deliverable: One-page personal brand strategy with a content calendar
Deployment & Capstone Project
Week 6Shipping a production AI application accessible to the world.
- FastAPI: routes, request/response models, middleware, async handlers
- Containerisation with Docker: Dockerfiles, images and layers
- VPS setup: provisioning, SSH, firewall and process management
- Reverse proxying with Nginx: routing, SSL and subdomains
- Connecting a frontend UI to a deployed AI backend
- Monitoring, logging, security: keys, rate limiting, CORS, sanitisation
Presenting Your Work & Career Acceleration
- Demo storytelling for a non-technical audience
- The capstone structure: problem, solution, demo, impact, next
- Negotiating your value: salary, rates and scope
- Your 90-day post-cohort action plan and alumni network
Deliverable: Capstone demo presentation + 90-day post-cohort action plan
Meet your instructors
Mayowa Opeyemi AnimasaunFounder
Founder & Lead Developer · Lead Instructor
Mayowa founded Sephar-Innovations and leads its engineering and AI training. A practitioner first, he builds production AI systems and software for real clients — and teaches the same hands-on skills he uses every day. He personally mentors cohorts from first principles to a shipped, deployed project.
Toluwanimi Odunewu
AI Engineer
Toluwanimi Odunewu is an AI Engineer at Sephar Innovations, where he builds the AI-driven features for products designed and engineered by the company. A graduate of the University of Lagos who taught himself machine learning and software engineering, he specializes in agentic systems built on large language models, turning AI into reliable, production-ready tools. Outside Sephar, he builds open-source AI for scientific research and leads AiSynth Labs.
Ready to start building?
Reserve your seat in the next cohort, or enrol your team. We'll issue a formal invoice and you can pay online or by bank transfer.