10 yrs · 10M+ users · 20 countries · $15M+ saved · 12+ releases · Open to work
| Project | What it does | Business logic | Frameworks / Stack | Live | Status |
|---|---|---|---|---|---|
| pm-os | PRD → full SDLC via multi-agent orchestration | Cut PRD-to-eval-ready cycle from weeks to hours | Claude Skills · MetaGPT · Promptfoo · Python · FastAPI | ↗ | v1.0 |
| MarkIQ Capital AI | AI capital-intelligence terminal (embedded app + assistant) | Read-only decision support; no autonomous trading | React · FastAPI · yfinance · multi-provider AI router (Claude · GPT-4o · Gemini) · Recharts | ↗ | v2.0 |
| Agentic Compliance Evals | HITL eval harness for regulated AI (KYC · AML · Sanctions · CDS · FHIR) | Gate deployment on calibration, not accuracy | Promptfoo · YAML · GitHub Actions CI · Claude · GPT-4o · Gemini · MIT | — | v0.3.0 |
| AI EHR · BeBuddy AI | Private wellness PWA (stress · sleep · mood · routines) | On-device only, no medical advice, no data leaves the phone | PWA · Service Worker · IndexedDB · Vanilla JS · Manifest | ↗ | bebuddy v1.0 |
| ai-pdf | Full AI PDF suite (edit · scan · OCR · merge · sign · chat) | Privacy-first: 100% client-side; AI features opt-in via proxy | PDF.js · pdf-lib · Tesseract WASM · OpenCV · jscanify · FastAPI multi-provider router | ↗ | v1.0.0 |
| ResumeForge | Resume + JD → ATS-optimised 2-page Word doc | 90% of resume gain = extracting context the candidate forgot | Python · Streamlit · FastAPI · python-docx · Gemini free tier / Groq / Ollama | ↗ | POC |
| prd-templates | MIT PRD templates for AI-native teams | Forces regulator citation + HITL spec + eval plan on every PRD | Markdown · fork-friendly · MIT | — | v1.0.0 |
| B2C Meta Intelligence | UK adtech + sports meta-intelligence (any sports property) | Ready for FIFA World Cup 2026 · Premier League · NFL Super Bowl · IPL · F1 · Olympics | Markdown research · BARB · UKOM · Nielsen · WARC · Ofcom/ASA-aligned | — | v1.0 |
| Portfolio Version 2 | Migrating case studies to Problem → Approach → Trade-offs → Outcome → Postmortem | Every case reads as a real product decision, not marketing | Markdown · Mermaid · GitHub Pages · Cloudflare Pages | ↗ | ongoing |
🛠 pm-os — Latest MVP
A PM operating system: turn a PRD into a full SDLC by orchestrating Claude Skills · MetaGPT · Promptfoo. Built by a Senior PM for regulated AI.
flowchart LR
PRD[PRD.md] --> S1[Claude Skill<br/>prd-to-sdlc]
S1 --> FRD[FRD + SRS]
FRD --> S2[MetaGPT<br/>architecture + code]
S2 --> Eval[Promptfoo<br/>eval gates]
Eval --> Release[Release-ready SDLC]
🌐 Live app · 📖 Repo · 🧠 prd-to-sdlc skill · ⚡ y-score-readiness · 🏷 v1.0
Agentic AI onboarding — 7 bounded agents · deterministic gates · event-sourced audit log. Mapped to FATF · EU AI Act · GDPR · BSA · Basel III.
flowchart LR
A[Customer/Teller] --> B[Intake]
B --> C[Doc Intel]
C --> D[KYC Validation]
D --> E[Sanctions / PEP]
E --> F[AML Risk]
F --> G{Risk Gate}
G -- Low --> H[Auto-Approve]
G -- Med/High --> I[Compliance Review]
H --> K[CRM / Core Banking]
I --> K
F --> L[(Audit Log)]
📖 Case study · 🏗 Agent design · ⚖️ Compliance · 🌐 Live page
🧠 Capital AI IQ — AI Capital Intelligence Terminal
v2 upgraded: Bloomberg-style terminal UI · embedded live app on the showcase page · sidebar AI assistant for symbol-level Q&A · multi-provider AI router (Claude · GPT-4o · Gemini) · risk / opportunity / uncertainty scoring · price chart + news signals · 30+ curated securities plus any yfinance-supported symbol · explicit guardrails (read-only decision support; no broker execution; no autonomous trading).
flowchart LR
M[Market Data<br/>yfinance] --> B[FastAPI Backend]
N[News RSS] --> B
B --> R[Opportunity / Risk / Uncertainty]
R --> AI[Free AI Router<br/>Claude / GPT-4o / Gemini]
AI --> I[Insight Bullets]
B --> UI[React Dashboard]
I --> UI
UI --> H[Human Decision]
🌐 Live showcase · 📖 Repo · 🖥️ Full app (cold-start) ·
📝 ResumeForge — AI resume-optimisation POC
Turns a master resume and a job description into an ATS-optimised two-page Word document. Built on the insight that most quality gain comes from extracting the context a candidate forgot to include — team size, real key-performance-indicator numbers, technical trade-offs, customer-discovery scope. A five-question discovery loop, tuned to the job description, does that extraction before generation.
flowchart LR
R[Master resume] --> D[Discovery loop<br/>5 questions, JD-tuned]
J[Job description] --> P[JD parser + gap analyzer]
D --> C[Enriched context]
P --> C
C --> G[Resume generator<br/>Claude / Gemini / Groq]
G --> W[ATS-scored 2-page Word doc]
🌐 Live application · 📖 Repository · 📘 PRD · 📋 MVP plan
📄 ai4pdf — Full-stack AI PDF tool
Edit · Scan · OCR · Merge · Split · Sign · Annotate · Redact · Fill forms · Chat/Summarize. 100% client-side (PDF.js · pdf-lib · Tesseract · OpenCV · jscanify). AI features via multi-provider router (Claude · GPT-4o · Gemini) through FastAPI backend.
flowchart LR
U[User drops PDF] --> C[Client-side PDF.js]
C --> Ops{Op}
Ops -- edit/merge --> L[pdf-lib]
Ops -- OCR --> T[Tesseract WASM]
Ops -- scan --> V[OpenCV + jscanify]
Ops -- chat --> A[AI router<br/>Claude · GPT-4o · Gemini]
L & T & V & A --> Out[Download / continue]
🌐 Live app · 📖 Repo · 🩺 Backend health · 🏷 v1.0.0
💚 BeBuddy AI — Private wellness PWA
Free, private wellness buddy — stress · sleep · urges · mood · routines · daily wins. On-device PWA, installable, works offline. Explicit non-medical framing.
🌐 Live app · 📖 Repo · 🏷 bebuddy-v1.0
More sample case studies (EHR · Contract Ingestion · US Health · UK Consumer)
| Sample | What it demonstrates | Stack |
|---|---|---|
| BeBuddy AI (wellness) + EHR platform | Live private wellness PWA + healthcare reporting/CDS layer · Repo | PWA · FHIR · HL7 · HIPAA-aware · on-device |
| Gen AI Financial IQ | Agentic banking contract ingestion | RAG · Banking · Contract intel |
| US Health Data Strategy | Archival economics, HL7 retrieval | Snowflake · Lineage |
| B2C Consumer Meta Intelligence | UK media intelligence | Consumer intel · UK |
| Company | Role · Reports to | Domain | Scope + Outcome |
|---|---|---|---|
| Nagarro · 2022–now | Staff Consultant · Head of Product | Enterprise SaaS agentic AI · global consulting productivity platform | Drain3 + Claude Opus RAG on AWS Bedrock · GPT-4o/Claude/Gemini multi-eval · RLHF + HITL · CSAT 88→94% · 1K→20K users in 3 months · $5M saved · $5M investment secured · delivery recognition Q4'25 |
| Nagarro | Sr PO · Head of Product | 0-to-1 healthcare SaaS · regulated health and financial wellness platform | HL7 · FHIR · ICD-10 · Apple Health + Google Fit · consent APIs · clinician-in-loop · 10M+ users · 20 countries · revenue +20% · claims -15% · 99.8% accuracy · 0 PII breaches |
| Nagarro | Sr PO | Humanitarian donor platform · Salesforce NPSP | 4 markets · Stripe+Vindi · WhatsApp/SMS360 NLU · Einstein AI · Discovery in 1 month · 5 months ahead |
| Cognizant · 2021–22 | Team Leader · Product Director | Enterprise conversational AI · 50 contact centres | Dialogflow + CCAI + OpenAI · 24-person team · 250+ user stories · SAFe PI · $10M+/yr saved · CSAT +35% · AHT -25% · intent recall 98% · containment +42% · latency 10s→7s |
| C2B Tech · 2020–21 | BA · PO | 0-to-1 insurance · largest South African bank | ICD claims · provider network · ERPNext · 30+ interviews · 120-page BRD · 95% MVP on-time · GBP 2M expansion · underwriting -45% |
| ICICI Securities · 2017–19 | Assistant PM | Salesforce Capital Markets · BFSI | SIP fund lifecycle · Stripe · SFTP recon · SMS360+WhatsApp · risk-profile engine · recon 40→80% · 12 releases 0 rollbacks · 200+ staff trained |
| The Stone Palace · 2015–17 | BA | B2B industrial fleet analytics | Lazy-loading arch · SLA/SLO · Tableau · Waterfall→Kanban · Efficiency +40% · SLA breaches -55% |
Also: Aviation loyalty mobile UX (RFP win · WCAG mobile UX) · industrial service fleet analytics (RFP win · PRD for AI recommendation engine)
Start with the business outcome. If it doesn't ladder to a measurable outcome, it's documentation, not product work. Convert ambiguity into structured artifacts. PRDs · BRDs · FRDs · ERDs · sequence diagrams · decision logs. Design for governance before scale breaks trust. Compliance, audit, HITL are first-class. Retrofit is 10× more expensive. Measure AI, don't demo it. Golden test sets · eval pipelines · RLHF guardrails · drift monitoring · AI observability. Keep architecture understandable for executives, engineers, and operations alike.
flowchart LR
A[Business Outcome] --> B[Stakeholder Pain]
B --> C[Structured Artifacts]
C --> D[Governance First]
D --> E[Measurable AI]
E --> F[Understandable Arch]
F --> G[Risk Register]
📘 decisions.md · 📊 metrics.md · 📄 CV
🤖 AI Product Platforms — Agentic AI · Deep RAG · Drain3+LLM hybrid · GPT-4o/Claude/Gemini multi-criteria eval · MCP · A2A · LangGraph · Semantic Kernel · RLHF · guardrails · drift monitoring · AI observability · golden tests · HITL.
📊 Data & Analytics — Snowflake · Kafka · dbt · Airflow · Chroma DB · data contracts · lineage · Power BI · Tableau · Kibana · Grafana · Amplitude · Mixpanel · A/B · DAU/MAU · cohort · North Star + OKRs.
🏥 Healthcare — HL7 · FHIR · ICD-10 · EHR · consent APIs · hospital provider networks · claims · copay · care pathways · clinical CDS · clinician-in-loop · patient stratification · NCPDP · HIPAA.
🏦 BFSI & Fintech — Payments (Stripe · Vindi) · reconciliation · onboarding · KYC · AML · sanctions/PEP · capital markets · SIP · fund lifecycle · risk profiling · Salesforce Capital Markets · Salesforce NPSP.
🧭 Strategy & Delivery — Vision · roadmap · market sizing · competitive analysis · RICE · MoSCoW · OKRs · RFP · GTM · SAFe PI · CSPO · JTBD · story mapping · 12+ production releases zero rollbacks.
All 17 certifications (click to expand)
| Area | Credential | Issuer | Evidence |
|---|---|---|---|
| Agentic AI | Building Intelligent AI Agents | Pearson/O'Reilly | Credly |
| LLM APIs | Claude with the Anthropic API | Anthropic | |
| AI Coding | Claude Code Foundations | Anthropic | 2026 · verify |
| Generative AI | Foundations in Generative AI | IBM SkillsBuild | certificate |
| Generative AI | Introduction to LLMs | IBM SkillsBuild | Digital sticker |
| Responsible AI | Ethical Considerations for GenAI | IBM SkillsBuild | Digital sticker |
| Product Analytics | Product Analytics Certification | Product School | Jan 2026 · cert_5q25t41p |
| Product Ownership | Scrum Product Owner Course | Pearson/O'Reilly | Credly |
| AI Product | AI Product Management Course | Simplilearn | certificate |
| AI Product | AI in Product Management | O'Reilly | 2025 |
| Security | AI Product Security | Dec 2025 | |
| Agile | CSPO | Nagarro | 2025 |
| PM | Technical Product Manager | Nagarro | 2025 |
| PM | Diploma in Product Management | Alison | 2024 |
| Data | Data Analytics | Simplilearn | 2023 · 4473790 |
| Product Ownership | Robust Scrum Product Owner | Udemy | 2023 |
| Education | Verified International Academic Qualifications | WES | Credly |
Tools & tech stack (click to expand)
Product — Jira · Confluence · Aha · Azure DevOps · Miro · Figma · Balsamiq · Lucidchart · ProductPlan · MS Visio · Rally Engineering & APIs — Postman · Git · Docker · Terraform · AWS Bedrock · Azure · GCP · Kafka · REST · Webhooks · APIM · OAuth 2.0 · PostgreSQL · Python · JSON · XML · SFTP AI & Integration — OpenAI · Anthropic · Azure OpenAI · Gemini CLI · LangChain · LangGraph · Playwright · Semantic Kernel · CCAI · Dialogflow · Salesforce · Salesforce Einstein · Stripe · Vindi · ERPNext Data & BI — Snowflake · Power BI · Tableau · Kibana · Grafana · ELK · Amplitude · Mixpanel · Google Analytics · dbt · Airflow · Chroma DB Healthcare & Compliance — FHIR tools · HL7 validators · EHR connectors · HIPAA logging · GDPR · PCI-DSS · consent APIs · ICD-10 · NCPDP · FATF · EU AI Act · BSA/CIP · X12 EDI
- Post Graduate Certificate, International Project Management — Anglia Ruskin University, London 🇬🇧 · 2013–2014
- Bachelor of Engineering, Computer Science & Engineering — First Class — Anna University, Chennai 🇮🇳 · 2008–2012
Open to work · IST (UTC+5:30)
🌐 vittobh.github.io · 📅 Book 15-min · 💼 LinkedIn · 📧 Email · 📄 decisions.md · 📊 metrics.md
Product Manager · Senior PO · AI Product Manager · Healthcare Data PM · BFSI Fintech PM · AI Platform PM · Enterprise SaaS leadership.
Currently reporting to Head of Product / Product Director · 10 yrs · Fortune 500, Big Three consulting, global enterprise clients.
📊 Reproducible results — agentic-compliance-evals
Latest CI-graded run across 15 cases × 3 frontier models — full scorecard:
| Model | Correctness | Grounding | Calibration | Safety | Overall |
|---|---|---|---|---|---|
| claude-3-5-sonnet | 87% | 82% | 79% | 96% | 86% |
| gpt-4o | 84% | 76% | 71% | 91% | 81% |
| gemini-1.5-flash | 71% | 64% | 63% | 88% | 71% |
Finding: every frontier model fails calibration before correctness. Gate on calibration, not accuracy. An 87%-accurate model with 79% calibration is more dangerous than a 71%-accurate model with 88% calibration — the first is wrong with conviction; the second tells you when to escalate.
Reproduce: README · PRD · ARCHITECTURE
- Client Performer Award Q4 2025 — Nagarro · Enterprise SaaS Agentic AI platform for a Big 3 US Management Consulting Firm.
- Writing — Two teardowns published on regulated AI, cadence every 2 weeks. /writing/.
- Open source — Author/maintainer of pm-os, agentic-compliance-evals, ai-pdf, prd-templates.
- Speaking — Available for meetups and podcasts on regulated-AI product management. Book a topic conversation.
Full detail: awards.md · case-studies.md · hire-me.md.
About me — full narrative
Product Manager and Senior Product Owner with 10 years leading product strategy, roadmap, and delivery for B2B and B2C SaaS, Mobile, and Web platforms across BFSI, healthcare, telecom, and enterprise consulting.
Drives executive decisions from discovery to launch, builds cross-functional teams aligned to product vision, and owns the full lifecycle: stakeholder interviews, PRD, BRD, FRD, user stories, RICE prioritisation, North Star metrics, and OKR tracking.
Delivers measurable outcomes across CSAT, retention, conversion, AHT, containment rate, reconciliation accuracy, precision, recall, and cost reduction through AI integration, A/B testing, and KPI diagnostics that validate every product decision.
Shipped platforms to 10 million users across 20 countries for Fortune 500, Big Three consulting, and global enterprise clients.
Work sits at the intersection of product strategy, data architecture, and AI platform design. Strongest where the problem is messy: many systems, many stakeholders, high compliance pressure, unclear workflows.
Led 0-to-1 product builds, authored 120+ page BRDs, written 250+ user stories in a single engagement, created 30+ wireframes per project, and supported 12 production releases with zero rollbacks.
Care about trusted AI in real enterprise settings: grounded retrieval, human-in-the-loop validation, audit trails, explainability, observability, latency control, hallucination reduction, measurable business outcomes.
Quick Stats — 6-cell grid
| 10 yrs | 20K | 10M+ | 20 | $15M+ | 12 |
| PM experience | enterprise users | consumers | countries | savings delivered | releases · 0 rollbacks |
Extended contact & artefacts
- 📄 Resume — PDF
- 🌐 Portfolio
- 📅 Book 15-min chat
- 📧 vittobha2025@gmail.com
- 📘 decisions.md — public ADR log
- 📊 metrics.md — portable 10-yr outcomes
"What I work on" — extended focus paragraphs
🤖 AI Product Platforms — Agentic AI · RAG · model evaluation · tool calling · MCP · A2A patterns · LangGraph-style orchestration · hallucination guardrails · AI observability · golden test sets · human validation workflows. Multi-criteria model evaluation across GPT-4o, Claude, Gemini on latency/cost/accuracy. Drain3 + LLM hybrid architectures on AWS Bedrock. RLHF guardrails, drift monitoring, AI observability with traces and spans, and human-in-the-loop governance.
📊 Data Products & Intelligence Platforms — Snowflake · metadata quality · lineage · audit trails · data contracts · canonical data models · retrieval strategy · data archival economics · executive KPI design. Data-pipeline design end-to-end from EHR/CRM/ERP ingestion to consumer-facing suggestion delivery. Schema validation and ML quality gates.
🏥 Healthcare Platforms — FHIR · HL7 · EHR integration · ICD-10 · clinical decision support · care pathways · consent management APIs · HIPAA-aware workflows · patient journey modeling · clinician-in-the-loop sign-off · NCPDP · hospital provider network management · claims processing · copay/underwriting rules.
🏦 BFSI & Fintech Products — Payments (Stripe, Vindi) · reconciliation (SFTP/XML/JSON) · onboarding · KYC · AML-aware workflows · capital markets · SIP journeys · fund allocation · exception triage · mobile banking authentication · agreement data models · risk-profile scoring engines · sanctions/PEP screening · FATF, EU AI Act, GDPR, BSA/CIP compliance mapping.
🧭 Product Strategy & Delivery — Product vision · roadmap · market research · competitive analysis · RICE · MoSCoW · OKRs · North Star metrics · RFP artifacts · release planning · SAFe PI planning · backlog management · stakeholder alignment across insurer-provider-user journeys · GTM planning with sales, marketing, and customer success · CSPO practices · JTBD · story mapping · 250+ user stories in a single engagement · 30+ wireframes per project · 120-page BRD authoring.
"How I think" — original 5 principles + rationale
flowchart LR
A[Business Outcome] --> B[Stakeholder Pain]
B --> C[Structured Artifacts<br/>PRD · ERD · Flows]
C --> D[Governance First<br/>Audit · HITL · Explain]
D --> E[Measurable AI<br/>Golden Tests · SLAs]
E --> F[Understandable<br/>Architecture]
F --> G[Risk Register<br/>not Wishlist]
- Start with the business outcome. If it doesn't ladder back to a measurable outcome, it's documentation, not product work.
- Convert ambiguity into structured artifacts. PRDs, ERDs, sequence diagrams, decision logs — make the invisible visible.
- Design for governance before scale breaks trust. Compliance, audit, human oversight are first-class requirements. Retrofit is 10× more expensive.
- Measure AI, don't demo it. Golden test sets, eval pipelines, confidence thresholds. Production AI without evaluation is theatre.
- Keep architecture understandable for executives, engineers, and operations teams alike.
Historical engagements roll-up
🌍 Nagarro-era engagements: Regulated health and financial wellness platform · humanitarian donor SaaS platform across 4 markets · global consulting productivity platform with 20K enterprise users
🌍 Cognizant-era engagements: Enterprise conversational AI programme · 50 contact centres · 24-person team · 250+ user stories · 15 A/B tests · SAFe PI planning · Dialogflow + CCAI + OpenAI NLP · ELK + Kibana observability
🌍 Direct-employer engagements: ICICI Securities · Salesforce Capital Markets · 3-in-1 integrated account platform · Stripe SIP · SFTP XML bank reconciliation · risk-profile scoring engine · 200+ staff trained · 4-country CRM
🌍 Additional project contributions: Aviation loyalty mobile app (RFP success · aviation loyalty platform · WCAG mobile UX · Android validation) · industrial service fleet analytics (RFP success · PRD for AI recommendation engine · RPM indicators · predictive maintenance · root cause analysis)
Full tools list — Product · Engineering · AI · Data · Healthcare
Product — Jira · Confluence · Aha · Azure DevOps · Miro · Figma · Balsamiq · Lucidchart · ProductPlan · MS Visio · Rally
Engineering & APIs — Postman · Git · Docker · Terraform · AWS Bedrock · Azure · GCP · Kafka · REST APIs · Webhooks · APIM · OAuth 2.0 · PostgreSQL · Python · JSON · XML · SFTP
AI & Integration — OpenAI · Anthropic · Azure OpenAI · Gemini CLI · LangChain · LangGraph · Playwright · Semantic Kernel · CCAI · Dialogflow · Salesforce · Salesforce Einstein · Stripe · Vindi · ERPNext
Data & BI — Snowflake · Power BI · Tableau · Kibana · Grafana · ELK · Amplitude · Mixpanel · Google Analytics · dbt · Airflow · Chroma DB
Healthcare & Compliance — FHIR tools · HL7 validators · EHR connectors · HIPAA logging · GDPR · PCI-DSS · consent management APIs · ICD-10 · NCPDP · FATF · EU AI Act · BSA/CIP · X12 EDI
Nagarro-era detailed scope (client names anonymized)
Global consulting productivity platform (Staff Consultant · Technical PM · reports to Head of Product)
- Defined enterprise SaaS agentic AI platform PRD, BRD, FRD; evaluated AWS Bedrock and Meta agentic AI orchestrator for production readiness
- Ran weekly Aternity crash correlation for Claude add-in cohorts; validated lineage with Anthropic architecture team
- Owned full RFP with market sizing, competitive analysis, and North Star metrics presented to C-suite
- Designed REST APIs, webhooks, event-driven microservices across Aternity, Riverbed, ServiceNow, and Snowflake as source
- Automated ServiceNow add-in approval workflows and Entra ID group provisioning
- Built RAG workflow pipeline with retrieval, re-ranking, and vector schema; tested in sandbox with Postman
- Architected Drain3 + Claude Opus RAG hybrid with AWS Bedrock managed endpoints; evaluated GPT-4o, Claude, Gemini on latency, cost, accuracy
- Built Office Health Check Pre-Log RCA for L1/L2 triage; set RLHF guardrails, drift monitoring, AI observability with traces and spans, and HITL governance
- Directed Azure-first SaaS for AI tool onboarding across Alpha, Beta, Production; automated ServiceNow intake-to-deployment with Entra ID provisioning
- Claude Opus AI-assisted intake with human-in-the-loop; managed Claude add-in stability across 20,000 enterprise consulting users
- Outcomes: CSAT 88→94% · accuracy 99% · tickets -40% · effort -70% · 1K→20K users in 3 months · $5M saved Q3 2025 to Q2 2026 · crash resolution validated with architecture review · delivery recognition Q4 2025
Regulated health and financial wellness platform (Sr PO)
- 0-to-1 B2B Mobile and Web SaaS · Owned strategy from CDM POC to mobile launch in Singapore (1M users) expanding to 20 countries (10M users)
- Integrated financial wellness SDK analysing user income, expenses, and savings to recommend loans, investments, and rewards
- Authored PRD, BRD, FRD; defined GTM with sales, marketing, and customer success
- Designed REST API contracts, CMS (AEM), CDN (Akamai), FHIR/EHR data models, ICD-10 mapping, ER diagrams, BPMN workflows
- Integrated Apple Health and Google Fit SDKs for iOS and Android; Vidyo video and Google Maps clinic discovery
- Sandbox tested with partner providers and vendors through B2B gateway
- Built benefit eligibility, consent management, and cohort segmentation engine
- Defined data contracts, schema validation, and ML quality gates with clinician-in-the-loop sign-off
- Drove data lineage mapping, traceability, audit logs, and PII governance across CRM, ERP, and third-party integrations
- Outcomes: 10M users · 20 countries · multi-tenant/multi-currency/multi-language · revenue +20% · claims -15% · 99.8% data accuracy · HIPAA · GDPR · PCI-DSS with 0 PII breaches
Humanitarian donor SaaS platform — Salesforce NPSP (Sr PO)
- Web SaaS on Salesforce NPSP · Drove strategy, PRD, user stories, delivery across Brazil, Bulgaria, Greece, Romania
- Ran discovery workshops, vendor evaluations, and roadmap alignment across country programmes
- Integrated Stripe and Vindi payment gateways with automated bank reconciliation across 4 European banks
- Built WhatsApp and SMS360 flows with NLU for donor intent detection
- Defined AI segmentation and reactivation models via Salesforce Einstein
- Designed anomaly detection and fraud flagging logic
- Outcomes: Solved Ultra HNI and HNI donors from anonymous data with PII enrichment · enabled personalised appreciation campaigns driving recurring donations · 6 banks integrated across 7 countries with Pix Key, QR Code, and barcode reconciliation
Non-Nagarro-era detailed scope
Enterprise conversational AI programme (Cognizant, Team Leader · reports to Product Director)
- Established strategy for $10M+ omnichannel SaaS (Mobile, Web, Contact Center) across 50 contact centres
- 24-person team · 250+ user stories · mentored 3 product owners · managed backlog refinement and release train
- Selected Dialogflow CCAI + OpenAI NLP through multi-criteria evaluation
- Designed webhook integration, microservices, API orchestration, and omnichannel routing across mobile, web, and contact center
- Built ELK stack and Kibana observability dashboards for intent detection and agent performance
- Designed 15 A/B tests tracking CSAT, conversion, and retention
- Defined North Star metric (containment rate) and OKRs for the programme
- Reduced response latency from 10s → 7s through KPI-driven diagnostics
- Outcomes: CSAT +35% · AHT -25% · $10M+/yr saved · intent recall 98% · containment +42% · 8 production sprints, 0 missed go-lives
Large South African banking group — 0-to-1 insurance (C2B Tech, BA · PO)
- Designed onboarding flow converting existing bank users into insurance customers
- Built custom premium pricing engine using PII (age, income, location) and health history to generate real-time quotes
- Rules engine maps risk questionnaire answers to risk bands per health and motor insurance lines
- Automated policy PDF and payment via ERPNext
- Defined user stories, REST API contracts, and mobile SDK specs for iOS and Android
- 30+ stakeholder interviews using story mapping and JTBD
- Authored 120-page BRD · 30+ wireframes · ER diagrams · decision tables · validated rules in PostgreSQL
- Ran UAT across PHP-based CRM and benefit management platform
- Defined copay structures and underwriting rules; coordinated with insurance and payment partners for go-live
- Outcomes: 95% MVP on-time · GBP 2M expansion · completion +28% · underwriting -45% · manual effort -80% · 3 product lines in 8 months
ICICI Securities — Salesforce Capital Markets equity advisory (Assistant PM · Equity Advisor)
- Managed Salesforce equity advisory mobile & web app for ICICI 3-in-1 integrated account platform (savings, demat, trading as single identity)
- Full fund lifecycle: fund add, verify linked bank account, allocate to trade, settle, and release
- Stripe payment integration for SIP campaigns
- Automated bank reconciliation via SFTP XML file transfers with ERP partners
- Led analytics across DAU, MAU, retention, and conversion funnels
- Built SMS360 and WhatsApp automated alerts for every equity transaction (buy/sell confirmations, fund credits, SIP triggers)
- Drove CRM modernisation with Salesforce batch processing and cohort segmentation across 4 countries
- Built automated risk profile assessment (net worth, income source, investment timeline)
- Dynamically locked/unlocked products (F&O, small-cap funds restricted unless Aggressive profile)
- Profile mismatch guardrails halt transactions exceeding risk parameters
- Owned UAT for banking and trading systems · captured BFSI regulatory requirements
- Outcomes: Client acquisition 90% · retention 95% · pledge processing +20% · recon 40→80% · manual entry -70% · 12 releases 0 rollbacks · 200+ staff trained
The Stone Palace — B2B Industrial Analytics (BA)
- Built from scratch for industrial machinery fleet analytics across 3 branches
- Real-time RPM, compressor performance, and maintenance alerts vs 7-year and 3-year historical baselines to flag early degradation
- Standardising requirement checklists via stakeholder workshops
- Solved large-data challenge by designing architecture with lazy loading and click-to-call API pattern so engineers could pull years of fleet telemetry without crashing the browser
- Defined SLA targets, Tableau KPI dashboards, and role-based access
- Transitioned team from Waterfall to Agile-Kanban
- Outcomes: Efficiency +40% · onboarding -60% · SLA breaches -55% · real-time visibility across 3 branches · CAD-to-ERP invoice errors eliminated
Extended engagement note — additional RFP wins
Aviation loyalty mobile app
- RFP success · aviation loyalty platform with AI-driven rewards and WCAG mobile UX
- Supported requirement gathering and Android validation
Industrial service fleet analytics
- RFP success · PRD for AI recommendation engine
- Covering RPM indicators, predictive maintenance, and root cause analysis
Public-safe sample repos. Repositories above are public-safe demonstrations of methods, artifacts, and forward-deployed skills applied to production work under NDA. Actual client scope, systems, and metrics are on the CV — not in the repos.









