★ Featured service

AI Training
that actually lands

For teams who need to use AI — not just talk about it.

Four tiers of practical AI capability building, designed for regulated industries where "move fast and break things" is a sackable offence. Built from 20+ years in financial services and an MA in AI Ethics & Society. No vendor pitches. No hype. Just what your team needs to ship.

Book a discovery call
سلسلة
"An unbroken chain — from understanding to application."
Silsila / Training 2026
01 / Why Silsila

Training that respects
what your team already knows

Most AI training is either a vendor demo dressed up as education, or a 101 that treats senior professionals like beginners. Silsila programmes assume your team is technically literate and time-poor — and start from the problems they actually face on Monday morning.
/ 01
Built from delivery, not theory
Every module draws on real projects: collateral systems, repo trading, agentic workflows, regulatory change. Examples your team will recognise; problems they'll relate to.
/ 02
Ethics integrated, not bolted on
EU AI Act, PRA SS1/23, DORA — treated as design constraints, not afterthoughts. Grounded in academic work from an MA in AI Ethics & Society at Birkbeck.
/ 03
No vendor capture
Independent. No reseller agreements, no commissions, no platform allegiance. Your team learns the principles that survive whichever model wins this quarter.
02 / Formats

Four tiers — match the depth to the need

Programmes scale from a single workshop for a senior team, to multi-week cohorts, to executive 1:1s, to fully white-labelled curriculum for in-house L&D teams.
01
Workshops
Half-day · Full-day · On-site or remote
Concentrated sessions for intact teams — product, engineering, risk, compliance. Walk in with a problem; walk out with a working approach. Heavy on hands-on exercises; light on slideware.
Common topics: Prompt engineering for analysts · Building safe LLM pipelines · Agentic system design · RAG architecture for regulated data · Evaluation & guardrails · AI governance fundamentals
Duration 4 / 8 hours
Group size Up to 20
Best for Team kickoff
02
Multi-week Cohorts
4 – 8 weeks · Weekly live + async
Depth programmes that move teams from awareness to capability. Each cohort works on a real project from your business — not a sandbox exercise — with weekly live sessions, peer review, and structured assignments. Outcome: shipped artefacts, not certificates.
Programme tracks: Agentic Systems for BAs & PMs · LLM Engineering for Regulated Industries · AI for Product Leaders · AI Risk & Governance in Practice
Duration 4–8 weeks
Group size 6–12
Best for Capability shift
03
1:1 Executive Coaching
Monthly retainer · 60–90 minute sessions
Direct, confidential sessions for executives, founders, and senior product leaders who need to think clearly about AI strategy without a sales pitch. Vendor evaluation, roadmap pressure-testing, board-deck rehearsal, hiring strategy — whatever's on top of mind that week.
Typical clients: Heads of Product · Heads of Transformation · CTOs at financial firms · Founders fundraising on an AI thesis
Cadence 2× / month
Commitment 3-month min
Best for Strategic clarity
04
Curriculum-as-a-Service
White-label · For L&D teams & academies
Full curriculum design and ongoing content updates for in-house L&D teams who want to run AI training at scale without building it from scratch. Includes lesson plans, assessment frameworks, instructor notes, and quarterly content refresh as the field moves. Your brand, your trainers — Silsila underneath.
Deliverables: Full curriculum & learning objectives · Slide decks & instructor notes · Hands-on exercise repositories · Assessment rubrics · Quarterly refresh cycle
Engagement Annual
Modules 8–20
Best for Scale
03 / Curriculum sample

What's actually taught

A representative slice of modules. Every engagement is scoped to the team's existing level and the problems they're trying to solve — nothing is taught for the sake of completeness.
Foundations
How LLMs actually work
Transformers from first principles. Tokenisation, attention, training, inference cost. Demystifies the box so teams can reason about it honestly.
Foundations
Prompt engineering for working professionals
Beyond "write better prompts." Structured prompting patterns, chain-of-thought, few-shot design, when to use which technique — with finance-domain examples.
Architecture
RAG systems for regulated data
Retrieval design, chunking strategies, embedding choices, evaluation. With explicit treatment of data residency, audit trails, and PII boundaries.
Architecture
Designing agentic workflows
Tool selection, agent loops, human-in-the-loop checkpoints, failure recovery. When to use multi-agent vs single-agent. When not to use agents at all.
Practice
Evaluation & guardrails
How to actually know if your AI system is working. LLM-as-judge, deterministic evals, drift detection, red-teaming. Spectre and PersonaEngine as worked examples.
Practice
Vendor & build-vs-buy
A practical framework for AI procurement. What to ask vendors, what to ignore, how to evaluate model performance honestly, when to build instead.
Governance
AI risk frameworks in practice
EU AI Act, PRA SS1/23, DORA, ISO 42001 — as design constraints. How to translate regulatory text into product requirements your engineers can build against.
Governance
Ethics for product teams
A working vocabulary for the harder questions: when does helpful become manipulative? What does "human oversight" really require? Drawn from current academic research.
04 / Try it free

The Silsila AI Tutor

The same scenario-driven method I use in paid engagements, open for anyone to try. A free, self-serve curriculum: six tiers from how these models actually work to shipping them responsibly in regulated environments, with a scenario check after every lesson. No sign-up, progress saved on your own machine.
Free · No sign-up · 6 tiers

Learn to build and govern AI you can trust

Read it like an essay, or flip to a retro terminal mode and drive it from the keyboard. Either way it's the same curriculum: foundations, prompting, RAG, agents, evaluation, and governance, each grounded in a real scenario you'd meet on the job.

Open the tutor
05 / Who it's for

Built for teams who can't
afford to get this wrong

Start a conversation

Most engagements begin with a 30-minute discovery call — no slides, no proposals, just a conversation about what your team actually needs.

hello@silsiladigital.com