Lead Solution Architect · AI & Agentic Systems

Tusshar
Kalra.

Twenty-six years architecting banking transformation. Now building production agentic AI systems with my own hands.

Based in
Stockholm, Sweden
Work rights
Swedish citizen · Full EU
Availability
Open to senior roles - Q4 2026
About

The rare profile banks are hiring for right now.

Lead Solution Architect with 26 years in banking technology across Europe and Asia. My domain is regulated finance: KYC/AML, customer due diligence, corporate and private lending, credit origination and risk, digital onboarding, client lifecycle management. My platforms are PEGA, where I founded Swedbank's Centre of Excellence, and Appian, where I currently architect Länsförsäkringar Bank's lending transformation.

What makes this profile uncommon in 2026: I have independently designed, deployed, and operate a production multi-agent AI system running 24/7 on self-managed infrastructure. Not an experiment. Not a weekend project. Working production code that I wrote, secured, and run every day. Banks and consultancies are hiring for AI transformation leadership and most senior architects can only talk about it. I build it.

KYC / AML
Lending Transformation
Credit Origination
PEGA
Appian
REST / OpenAPI
Azure DevOps
Agentic AI
Multi-Agent Systems · OpenClaw · Hermes Agent
LLM Orchestration
Claude · Qwen · Kimi · DeepSeek · Mistral · Gemma
ACP · MCP
SAFe POPM
Centre of Excellence Design
Live · In production

Nexus - a production multi-agent AI system.

Designed, built, and operated solo since 2025. A five-agent multi-tier system running 24/7 on self-managed infrastructure, with a hardened Telegram gateway, cross-provider LLM fallback, and an autonomous closed-loop upgrade pipeline that completed its first live production upgrade in May 2026. The architecture banks need for vendor-independent agentic AI adoption.

5
Specialist agents - Newton (research swarm), George (coding), Hermes (trading intelligence), Leonardo (vision), plus Claude Code via ACP
18
Security controls across five concentric trust zones - perimeter, gateway, application, state, agent
24/7
Uptime on self-managed VPS infrastructure with Nginx + TLS, observability and audit trails
8
Scheduled workflows plus a 19-category nightly maintenance regime, with dual-delivery reporting (Notion + Telegram)
01
Five-agent multi-tier orchestration coordinated by Nexus as Tier 1 supervisor with intent classifier and capability-typed dispatch - Newton (swarm-capable research with parallel verifier children), George (coding), Hermes (trading intelligence), Leonardo (creative + deterministic render), with Claude Code integrated via Agent Communication Protocol. Five foundation-model families at primary tier (Qwen, Kimi, DeepSeek, Mistral/Devstral, Gemma) with two-deep cross-provider fallback - zero single-vendor lock-in.
02
Eighteen security controls across five concentric trust zones - perimeter (UFW deny-incoming, fail2ban, SSH key-only), gateway (loopback-only bind, token auth, prompt-injection detection, Telegram allowlist, DM-only enforcement), application (per-agent session isolation, denied-tool policies for small models), state (mode-600 secrets, no hardcoded credentials, nightly integrity checks), and agent (paid-API fallback discipline, observer-based daily spend hard cap).
03
Claude Opus 4.7 integrated via ACP protocol for autonomous coding tasks delegated from the orchestrator without manual handoff - the production pattern for agent-to-agent collaboration across foundation models.
04
Operational tooling built in: YouTube transcript ingestion, PDF reading, real-time market data, a 4-layer search stack (Tavily, DuckDuckGo, Ollama Web, Serper), and automated report generation.
04
MCP-native integrations across Notion direct API, SuperMemory, Groq Whisper, Gmail and Calendar OAuth. Browser stack for SPA-gated regulatory and careers portals - Playwright MCP plus system Chrome plus patchright stealth, accessibility-tree-first snapshots. Brand-consistent PDF delivery via WeasyPrint with zero LLM in the render path.
05
Eight scheduled workflows plus a 19-category nightly maintenance regime - task hygiene, database maintenance, agent health, credentials, MCP health, browser stack, cost tracking, VPS OS, plus a meta-monitor watching the regime itself. Six audit tables across agent health, model failovers, cron runs, upgrade history, active tasks, and retry attempts. Cron LLM hygiene refactor reclaimed ~132k LLM-seconds per week.
05
Two full security audits completed resolving 40+ findings - .env hardening, gateway token isolation, sandbox configuration for small models, plaintext-secret removal, CVE remediation, fail2ban, and UFW lockdown.
Case study 01 · Security

A Telegram channel, hardened like a production banking gateway.

Eighteen layers of control - token isolation, signed-message verification, replay defence, rate limits, and an audit trail. The same hardening pattern bank procurement committees ask architects to apply to third-party chat channels.

Read case study →
Case study 02 · Orchestration

Multi-model orchestration without vendor lock-in.

Nexus routes across GLM, MiniMax, Kimi, Gemma, and Claude with a fallback chain that survives a provider outage mid-task. The vendor-independence pattern DORA Article 28 now asks banks to design for.

Read case study →
Case study 03 · AFC

Agentic KYC - unravelling corporate UBO across jurisdictions.

Enhanced CDD, adverse-media screening, and beneficial-ownership unwind under AMLR Article 42 and the BORIS register. Research swarms and vision agents as proof-of-build components - the AFC transformation shape banks are budgeting for in 2026.

Read case study →
Case study 04 · Lending

Agentic AI across the consumer-lending lifecycle.

Origination through servicing - PSD2/3 data aggregation, document ingestion at application time, affordability beyond traditional credit scoring, real-time decisioning, back-book repricing. Same component-level proof-of-build anchoring under PSD3 and AI Act Annex III.

Read case study →
Case study 05 · Autonomy

Cron-Driven Autonomy

Eleven scheduled jobs that keep a multi-agent system honest while nobody is watching. The overnight batch reimagined as an agent layer under DORA Art. 17.

Read case study →
Case study 06 · Oversight

When the watcher couldn't be trusted.

A model-failover watcher that flipped a healthy system onto a fallback in a tight loop - because the cloud was rate-limiting, not failing. The boundary between monitoring and decision-making under AI Act Article 14.

Read case study →
Case study 07 · Governance

The bleed - AI compute as a treasury function.

Two paid-API leaks caught by a balance poller on a daily timer. The forensics, the two rules that fell out, the observer-based circuit breaker that enforces them, and the case for treating production agentic compute as a treasury function.

Read case study →
Career at a Glance

Twenty-six years in banking transformation.

2023 - Present
Lead Solution Architect - Lending
Länsförsäkringar Bank, Stockholm
Sole architect for the Lending domain, owning the end-to-end architecture of NOVA - the bank's new Credit Origination System on Appian, replacing a 1998-era legacy platform and sized for 200,000 annual transactions across seven lending products. Delivered MLR3 and MLR4 to production; currently architecting the Appian containerisation and cloud migration. Authored v2.0 of the NOVA Non-Functional Requirements specification (ISO/IEC 25010, GDPR, PCI-DSS, DORA, EBA).
2020 - 2023
Senior Solutions Architect - Anti-Financial Crime
Swedbank, Stockholm
Led Swedbank's enterprise KYC transformation on PEGA. Delivered the corporate customer segment on schedule. Owned the complete integration model.
2018 - 2020
PEGA Centre of Excellence Founder & Product Owner
Swedbank, Stockholm
Founded the CoE from scratch - operating model, governance, and architectural standards adopted across all PEGA product teams. Designed and introduced the enterprise Situational Layer Cake architecture. Led the CoE as Product Owner (Certified SAFe POPM). Architected cloud migration strategy.
2016 - 2018
PEGA Delivery Lead
Capgemini / Swedbank, Stockholm
Delivered the Future Lending Platform and Credit Origination System end-to-end on PEGA across private and corporate credit product streams.
2013 - 2016
Delivery Manager
Wipro Technologies, Bangalore
Lloyds TSB Integrated Credit Risk System and ANZ banking programmes.
2012 - 2013
BPM Project Manager
Virtusa Consulting Services, Hyderabad
BNP Paribas PEMSA payment exceptions platform and Corner Banca case and workflow management solution.
2007 - 2012
IT Project Manager
HSBC Software Development, Pune
HSBC Invoice Finance, Receivable Finance, and Loan Applications - five years of global banking project delivery.
2000 - 2007
Project Lead
MphasiS (HP Company), Bangalore
Seven years foundation in banking project leadership and SDLC.
Contact

Open to senior roles in Q2 2026.

Lead Solution Architect, Principal Architect, AI Transformation Lead, or Head of AI Engineering roles - at banks, fintechs, or consultancies in English-working environments. Permanent or contract. Stockholm preferred; open to relocation across the EU, UK, Switzerland, Singapore, and Dubai for the right role.

Email
LinkedIn
Phone
+46 72 240 3490