Warsaw · ING Hubs Poland · FinCrime × AI since 2023

Kiran
Mohan

_

The rare professional who bridges 8+ years in the AML/TM field
with hands-on Agentic AI deployment.

CTMA · ACAMS DeepLearning.ai Agentic AI 38 Jurisdictions MSc Aerospace · Warsaw TU ING · Citibank · MoneyGram
Scroll
AI
Core
The Differentiator

Where compliance depth

meets agentic intelligence.

Most AML experts can't architect multi-agent systems. Most AI engineers don't know a SAR from a STR. Kiran is the bridge — architecting production Agentic AI workflows and Copilot agents at ING, backed by a flawless record of zero critical audit findings.

0
Years AML/TM expertise
0
Jurisdictions covered
0
Production AI systems
0
Critical Audit Findings
🏆
CTMA · ACAMS  +  DeepLearning.ai Agentic AI Certified  ·  4.3 GPA M.Sc Aerospace Engineering
The only credential stack spanning regulatory compliance AND multi-agent AI systems — grounded in engineering rigour
Career

8 years in

Financial Crime Prevention Industry.

01 / 05
Nov 2016 – Nov 2017
Citibank Europe PLC
Junior AML Analyst — Transaction Monitoring
  • Sanctions/AML escalations research; collaborated with Relationship Managers
  • Mantas scenario tuning and local regulatory adaptation
02 / 05
Apr 2019 – Nov 2020
MoneyGram International
Compliance Analyst III / II — Transaction Monitoring
  • Deputy Team Lead (10 FTEs); introduced workflow refinements boosting productivity by 15%
  • Authored & maintained process documentation, SOPs and training guides
03 / 05
Dec 2020 – Dec 2021
ING Hubs Poland
Senior AML Investigator / Business Analyst — TM Optimisation
  • Automated daily MI packs and productivity scorecards using VBA & Python
  • Complex alert investigations with 98% quality score
  • Developed VBA macro auto-generating investigation reports (25 min savings per case)
04 / 05
Jan 2022 – Dec 2024
ING Hubs Poland
Business Analyst / AML Expert — Project Team / Strategy, Design & Risk
  • Co-created enterprise-wide TM Optimisation framework (Quantexa & in-house ML models)
  • Designed Power BI dashboards covering KRIs/KPIs for global MLROs
  • Orchestrated threshold strategy reset across 20+ Jurisdictions
  • Zero critical findings in internal audit of 2024
05 / 05
Jan 2025 – Present
ING Hubs Poland
Senior Expert / Financial Crime Prevention — TM Backtesting & Design
  • Pioneered ING's first Global TM Effectiveness & Efficiency Assessment across 38 markets
  • Designed Microsoft Copilot agents automating business analysis report generation
  • Spearheading vertical integration of Agentic AI into existing TM processes
  • Principal advisor to MLROs across multiple regions; lead 3-person Specialists team
Agentic AI · Portfolio

Systems

I Build.

47 purpose-built domain agents across multi-agent swarms, intelligence platforms, and frontier research — engineered to production standards.

0
Domain agents built
0
Total agent ecosystem
1.91B
Tokens processed (3 weeks)
37:1
Output / Input ratio · concise direction, extensive agent execution
Open Source · Solo Projects
Completed
Content Factory v2.3

A fully autonomous multi-agent pipeline — from raw idea to published article, with no human prompting between steps.

Researcher
Web scraping · Trend detection · Source ranking · Topic clustering
01 / Input
Analyzer
Epistemological framing · Angle selection · Popperian critique · Argument scaffold
02 / Frame
Writer
Naval-style compression · Multi-format drafts · Substack / LinkedIn / YouTube scripts
03 / Draft
Editor
Tone calibration · Hard-to-vary pruning · Quality gate · Multi-channel output routing
04 / Publish
Pipeline Outputs
Long-form Article
article-template.md · Writer → Editor
🐦
Twitter/X Thread
twitter-thread-template-v2.md · Thread Agent → Editor
💼
LinkedIn Post
linkedin-post-template.md · Writer → Editor
📰
Newsletter Section
newsletter-template.md · Writer → Editor
YouTube Script
youtube-script-template.md · Writer → Editor
🎙
Podcast Notes
podcast-notes-template.md · Writer → Editor
4
AI Agents
6
Output Formats
0
Human Prompts
View on GitHub
Work in Progress
Orbital Inspect
5-agent satellite insurance intelligence
Multi-model adversarial validation — Codex GPT 5.4 as auditor, Gemini 3.1 Pro as validator. SSE streaming, durable jobs, human-in-the-loop decisions with admin-gated overrides.
5 AgentsMulti-ModelJWT/RBACSSEOpenTelemetry
View on GitHub
Ongoing Research · GPD Framework
Agent Economy Fraud
Agent-to-agent transaction network analysis
Fraud patterns in agent economic networks, validated against 81,000 real blockchain transactions via Dune Analytics. Statistical mechanics frameworks for emergent fraud typologies.
81K TransactionsDune AnalyticsNetworkXStatistical Mechanics
View on GitHub
Ongoing Research · GPD Framework
Agent Statistical Mechanics
Theoretical framework for multi-agent dynamics
Applying statistical mechanics — partition functions, phase transitions, scaling laws — to understand emergent behaviour in multi-agent AI systems.
Scaling LawsPhase TransitionsAgent DynamicsarXiv Target
View on GitHub
Open Source · Collaborative Projects
Completed

AEGIS

Emergency · Multilingual AI Response

Valencia, 2024. 215 people died despite a 12.5-hour warning — because emergency alerts failed to cross the language barrier. AEGIS is an AI system built to eliminate that gap: real-time distress detection across 12+ languages, instant unit dispatch, and voice response in the caller's own tongue.

Built in collaboration with Chetan Bavdhankar · Mistral AI Hackathon 2026

🌍
12+ languages · Real-time detection
🤖
3 Mistral agents · Ingest → Analyze → Respond
🎙
ElevenLabs TTS · Voice in caller's language
🆘
AEGIS Emergency
● Online · Multilingual
✓ SAFELY EVACUATED
47 seconds
Response time · Unit 7 dispatched · Rescue confirmed
Work in Progress
In Active Development
DeepInspect V2
8-agent bridge infrastructure intelligence
Full-stack platform: FastAPI + React + PostgreSQL. 8 parallel risk agents, swarm synthesis, 11-criteria scoring, 9 EU rating systems. CQRS, circuit breakers, 203+ tests, Playwright E2E.
8 AgentsFastAPIReactPostgreSQL203+ TestsDocker CI/CD
View on GitHub
Creative Experiment · Work in Progress

Cinematic

Video Pipeline

Building a zero-cost cinematic AI video pipeline on Colab T4 GPU. The architecture combines CogVideoX-2B for visual generation, Kokoro TTS for voice narration, and MusicGen for atmospheric soundtracks, all orchestrated locally via Python and assembled using FFmpeg.

Targeting full end-to-end Knowledge Architect video production

🎥
CogVideoX-2B · Inference server
🎙
Kokoro TTS & MusicGen · Audio synthesis
🎞
Python & FFmpeg · Local Orchestration
Node 1: Visuals
CogVideoX-2B 🟩
Node 2: Audio Synthesis
Kokoro TTS 🟨
MusicGen 🟨
Node 3: Assembly
FFmpeg Orchestrator
Thought Leadership

Knowledge

Architect.

Where David Deutsch's epistemology meets agentic AI, the philosophy of knowledge, and the hard problem of thinking clearly in a world being remade by machines.

Peer-Reviewed Research
📄
ScienceDirect · Elsevier Engineering Science and Technology · 2022
Initial research on thermal decomposition of 98% concentrated hydrogen peroxide in thruster-like conditions

An experimental investigation into the thermal decomposition characteristics of highly concentrated (98%) hydrogen peroxide under typical thruster operating conditions, providing foundational data for its use as a green monopropellant.

Rocket Propulsion Green Propellants Hydrogen Peroxide Thermal Decomposition
Read on ScienceDirect
Featured Writing & Episodes
Finding a Problem Worth Solving: The Skill You Were Never Taught
A deep dive into identifying and framing problems that actually matter.
Finding a Problem Worth Solving: The Skill You Were Never Taught
You feel the tension.
The Articulation Engineering: How to Explain Complex Thoughts
A strategic breakdown of how to translate deep technical understanding into clear explanations.
The Articulation Engineering: How to Explain Complex Thoughts
You understand something deeply.
The Physics of Focus: Why You’re Actually Distracted
An exploration of the mechanics of attention and how to reclaim your focus.
The Compression Revolution: The Real Way AI Changes Your Brain
Are we getting dumber or smarter in the AI era? The question itself reveals a fundamental confusion about how intelligence and compression actually change your brain.
Recognition

Awards & Honors

🏆 Lion Role Model Award
Individual excellence and leadership qualities (2024)
🎯 Transaction Monitoring Hackathon Winner
Best idea to improve TM efficiency (2024)
🧡 Orange Heroes Award
Peer-nominated recognition (2024)
🥇 Golden Wings Award Nominee
Individual excellence - TM pillar (2023)
⭐ Area Director's Award
TM Optimisation Department (2023)
Capabilities

Skills & Expertise

🛡️
Domain Expertise
TM Optimisation AML Frameworks Risk & Controls Design Regulatory Compliance Entity Resolution Multi-Jurisdictional Compliance
🤖
LLM & Agentic AI
Multi-Agent Orchestration LLM Ops & Quality Gates RAG & Evidence Pipelines Multi-Model Strategies Claude / Anthropic Gemini / Google GPT / Codex / OpenAI Mistral Ollama / OpenRouter Prompt Engineering
📊
Data Analytics
Python SQL Power BI KPI / KRI Dashboards Quantexa
🧭
Leadership
Change Management Stakeholder Engagement Training Design Competency Frameworks Team Development Agile (ING Model)
✍️
Communication
Technical & Business Communication YouTube (5.3K+ subs) Substack Workshop Design Report Writing Public Speaking
Let's Build

The future of agentic intelligence
needs a different kind of builder.

Senior Expert roles · Principal Consultant · Solopreneur advisory on Agentic AI for regulated industries.

Open to co-creating the next generation of intelligent systems. Let's explore how we can collaborate.

KIRAN