Dashboard / Candidate

Tarun Joseph

Placed in Senior Product Manager by best fit (no role stated; scored higher on this rubric) · 11_tarun_joseph.pdf

Private: squad_3@pg27.mesaschool.co · no phone found

Applied-role score

45/100

#6 of 16 SPM applicants

PM score

32/100

SPM score

45/100

Current status

Analysed

Below shortlist (AI suggestion)

Your decision

Why they fit

Problem -> Change -> Business/Operational Result 20%
The CV presents repeated, clear examples connecting the candidate's interventions to substantial operational and business outcomes. Quantified metrics span cycle time reduction, cost savings, error and fraud reduction, and direct revenue growth across multiple companies.

Where they fall short

Independent Decision Ownership −18 pts
The candidate has held founder and VP-level roles demonstrating substantial seniority and responsibility, including shutting down a venture. However, the CV focuses heavily on technical solutions and feature delivery without explicitly detailing the trade-offs, consequential product decisions, or choices made under incomplete information.
Evidence-Led Prioritisation and Reversal −16 pts
Evidence for this criterion is not present in the CV. While the CV lists A/B testing in skills and shows data-driven model tuning, it lacks any concrete narrative where new evidence caused the candidate to stop, reverse, or materially alter an earlier product direction.
Creates Systems Others Run −12 pts
The candidate repeatedly built technical software systems and automated workflow pipelines, but there is no evidence of creating team operating rhythms, documentation standards, or repeatable operational frameworks that team members adopted as a standard way of working.

Rubric breakdown: PM

SPM (applied)PM

Criterion	Weight	Score	Weighted	AI reasoning	Evidence from CV
Problem -> Change -> Result	25%	5/5	25.0/25	The CV provides repeated, high-impact examples where the candidate identified specific operational or technical bottlenecks, deployed targeted product solutions, and drove substantial measurable improvements in accuracy, cost, and processing time.	“Engineered an Anti-Hallucination Document Extraction Pipeline: Designed an advanced RAG + SLM architecture to accurately parse complex multi-region hotel invoices, surging extraction accuracy from 70% to 98% and cutting token compute costs by 40%.” “Solved Unstructured Financial Docs with GenAI: Transformed a highly manual underwriting process into a 10-minute AI decision engine using RAG + distilled SLMs, surging credit decisioning speed by 85% and enabling same-day disbursals.” “Engineered a Localized Document Intelligence Pipeline: Deployed spatial parsing and SLMs tuned for regional anomalies to autonomously validate non-standard handwritten mandi receipts and vernacular trade ledgers, collapsing loan approval TAT from 2 weeks to under 3 hours.” “Built a background telemetry and RPA pipeline that autonomously resolved endpoint compliance gaps, VPN drops, and cache bloat, slashing MTTR from 4 hours to a 1-click 10-second fix.”
Evidence-Led Decisions	20%	0/5	0.0/20	Evidence is not present. While the CV mentions 'A/B Testing' in the skills section and cites business metrics after product launches, it contains no examples of decisions being changed, stopped, or reprioritized based on operational data, usage metrics, or post-mortems.	Evidence not present in the CV.
Builds Repeatable Ways of Working	20%	0/5	0.0/20	Evidence is not present. The CV details several external product features and technical architectures delivered for customers, but does not provide examples of creating internal templates, documentation systems, review processes, or operating frameworks adopted by others.	Evidence not present in the CV.
Carries Ambiguous Problems Through to Closure	20%	1/5	4.0/20	The CV shows indirect evidence of managing complex stakeholder situations to closure during a company shutdown. However, it lacks examples of taking a messy internal escalation or ambiguous operational failure across functions through resolution with post-fix follow-up.	“Scaled to Pre-Series A term-sheet discussions before executing a structured, ethical shutdown , migrating users and returning angel capital following the pandemic-related loss of a key co-founder.”
Gets to the Source Before Designing the Fix	15%	1/5	3.0/15	The CV offers indirect hints of domain-specific user understanding, such as handling handwritten receipts and degraded thermal printouts. However, it provides no direct evidence of observing, interviewing, or shadowing users to inform and adapt the solution design.	“Overcame Digital Literacy Barriers via Voice AI: Launched a multilingual conversational AI assistant (English, Hindi, Nepali) for self-serve KYC and origination across tier-3/4 markets, slashing CAC by 35% and reducing field-agent dependency.” “Engineered a High-Speed OCR Pipeline: Deployed an asynchronous computer vision engine (Tesseract, AWS Lambda) custom-trained on degraded thermal paper to parse 30-item distributor invoices in under 15 seconds.”
Total	100%		32	The candidate demonstrates exceptional strength in linking technical solutions to tangible business and operational results across multiple ventures. However, the CV lacks evidence for internal ways of working, evidence-based decision pivots, and direct firsthand user discovery as defined by the rubric.

Scoring scale

0: No evidence in the CV
1: Indirect or participation-only evidence
2: Partial: some elements of the description, weak link to outcome
3: One clear example matching the description
4: Clear example with a visible result, candidate's own action linked to it
5: Repeated, strong examples fully matching the description, with results

Weighted contribution = score ÷ 5 × weight. Every quote is checked against the CV text; if none can be verified, the score is capped at 1.

Evidence

Verified CV quotes used (SPM rubric)

“Scaled to Pre-Series A term-sheet discussions before executing a structured, ethical shutdown , migrating users and returning angel capital following the pandemic-related loss of a key co-founder.” · Independent Decision Ownership
“surging extraction accuracy from 70% to 98% and cutting token compute costs by 40%. Saved ~INR 2.4Cr/year in compute and manual review costs, directly protecting gross margin on the AI product tier.” · Problem -> Change -> Business/Operational Result
“Transformed a highly manual underwriting process into a 10-minute AI decision engine using RAG + distilled SLMs, surging credit decisioning speed by 85% and enabling same-day disbursals.” · Problem -> Change -> Business/Operational Result
“collapsing loan approval TAT from 2 weeks to under 3 hours. Reduced field verification costs by INR 80L/year while enabling 4x throughput on the same operational headcount.” · Problem -> Change -> Business/Operational Result
“Architected a deep learning OCR pipeline to parse, flatten, and cross-verify PAN cards and bank statements, plummeting identity fraud by 40% and pushing extraction accuracy to 95%.” · Problem -> Change -> Business/Operational Result
“Scaled to Pre-Series A term-sheet discussions before executing a structured, ethical shutdown , migrating users and returning angel capital following the pandemic-related loss of a key co-founder.” · Multi-Party Resolution Under Pressure

Outcomes stated in the CV

Enabled INR 18Cr+ in annual SaaS ARR from enterprise tier subscription upgrades at FAI.
Cut compute and manual review costs by ~INR 2.4Cr/year via invoice extraction pipeline at FAI.
Facilitated INR 300Cr+ in working capital disbursals connecting 20,000+ MSMEs to 30+ lenders at Supermoney.
Digitized and disbursed INR 500Cr+ in annualized agricultural credit across 3 agri-corridors at Jai Kisan.
Processed INR 1,000Cr+ in annualized invoice volume via an integrated Supply Chain Finance engine at Jai Kisan.
Reduced field verification costs by INR 80L/year and loan approval TAT from 2 weeks to under 3 hours at Jai Kisan.
Lowered identity fraud by 40% and prevented ~INR 2.5Cr in estimated fraud-related write-offs at Payme.
Drove a 25% increase in enterprise contract renewals and saved ~$2M+ annually per enterprise client account at Anakage.

Anonymised CV (exactly what the AI saw)

• • [PROFILE URL REDACTED]
PROFESSIONAL SUMMARY
AI-Driven Product Leader & Former Founder with 8+ years of experience scaling zero-to-one FinTech, Enterprise SaaS,
and Agri-tech ecosystems. Rooted in strong systems-level thinking from a mechanical engineering foundation, I
specialize in architecting complex credit marketplaces and robust data pipelines (RAG architectures, predictive ML,
computer vision) that bridge the gap between Fortune 500 enterprises and “credit-invisible” MSMEs. Adept at driving
localized, inclusive innovation , deploying vernacular NLP agents across Northern and North-Eastern tier-3 markets to
break digital literacy barriers and slash CAC. A strategic operator with a proven track record of solving high-stakes data
bottlenecks, monetizing advanced analytics, and facilitating multimillion-dollar capital flows with zero-friction onboarding.
SAMPLE ASSIGNMENTS / PROJECTS
Unified Lead Queue: https://tinyurl.com/3d6xvmzm
Loan Management System Prototype: https://amount-flag-69813754.figma.site
SKILLS
Product Strategy & Business: 0-to-1 Product Development, Go-to-Market (GTM) Strategy, B2B2C Marketplaces,
SaaS Monetization, Agile/Scrum Leadership, A/B Testing, Enterprise Contract Renewals, B2B Supply Chain Telemetry.
AI & Machine Learning: RAG Architectures, Small Language Models (SLMs), Predictive ML (XGBoost, Random
Forest), Natural Language Processing (NLP), Computer Vision / Advanced OCR (Tesseract), Spatial Parsing,
Behavioral ML Credit Scoring.
Technical Architecture & Infrastructure: Multi-Tenant System Architecture, ETL Data Pipelines, RESTful APIs &
SFTP, AWS (Lambda, EC2, S3), Python (Django), Node.js, React Native, PostgreSQL, Redis, Enterprise ERP
Integrations (SAP, Workday).
FinTech & Domain Expertise: Supply Chain Finance (SCF), Agri-Fintech, Tax Reconciliation (GSTR 2A/2B),
Alternative Credit Underwriting, MSME Digital Lending, Identity Fraud Prevention, ITSM & Shift-Left IT Automation.
Localization, UX & User Journeys: Multilingual Product Scaling (English, Hindi, Nepali), Conversational AI Agents,
Zero-Friction Video-KYC Funnels, Digital Adoption Platforms (DAP), Robotic Process Automation (RPA), Omnichannel
Chatbots.
PROFESSIONAL EXPERIENCE
Director of AI Products | FAI | Jan 2025 – Nov 2025
Helping Fortune 500 clients like McKinsey, AMEX, and Tata reclaim millions in global and domestic tax through intelligent automation.
• Built a Scalable Tax Reconciliation Engine: Directed the development of a multi-tenant B2B/B2C platform with
robust ETL pipelines (RESTful APIs, SFTP, EML parsing) to synchronize GST portal data and SAP ERP ledgers ,
processing ~$0.5B in GSTR 2A/2B datasets for 1,200+ enterprise clients at 100% compliance. Directly enabled
INR 18Cr+ in annual SaaS ARR from enterprise tier subscription upgrades and expanded seat licenses.
• Engineered an Anti-Hallucination Document Extraction Pipeline: Designed an advanced RAG + SLM
architecture to accurately parse complex multi-region hotel invoices, surging extraction accuracy from 70% to 98%
and cutting token compute costs by 40%. Saved ~INR 2.4Cr/year in compute and manual review costs,
directly protecting gross margin on the AI product tier.
• Deployed Event-Driven CRM Automation: Built NLP-triggered autonomous vendor/client follow-up workflows,
reducing end-to-end pipeline TAT by 7 days. Accelerated revenue recognition by ~INR 4Cr/quarter by
unblocking delayed GST data collection that was stalling client billing cycles.
• Launched an Interactive BI Analytics Dashboard: Structured flattened relational data into dynamic GST/VAT
visualization modules with automated reporting for Fortune 500 stakeholders. Drove a 15% increase in product
revenue (~INR 6Cr incremental ARR) by proving ROI and accelerating upsell conversations with CFO-level
buyers.
External Product Consultant | Supermoney | Mar 2024 – Jan 2025
• Built a Full-Spectrum AI Credit Marketplace: Architected a dual-engine lending platform merging embedded
supply chain telemetry with alternative startup cash-flow data. Connected 20,000+ credit-invisible MSMEs and
fast-growing startups to 30+ lenders, facilitating INR 300Cr+ in working capital disbursals. Generated ~INR 3Cr in
platform origination fee revenue (1% average take-rate on facilitated loan volume) within the first 9 months.
• Solved Unstructured Financial Docs with GenAI: Transformed a highly manual underwriting process into a
10-minute AI decision engine using RAG + distilled SLMs, surging credit decisioning speed by 85% and enabling
same-day disbursals. Reduced per-loan processing cost by ~60%, unlocking ~INR 1.2Cr in annual
operational savings and enabling the team to scale from 200 to 800+ monthly disbursals.
• Built a Predictive Risk Brain for Dynamic Underwriting: Engineered ML models analyzing real-time alternative
data (seasonal FMCG transactions, supply chain signals) to dynamically throttle credit limits (INR 5L–INR 2Cr),
maintaining a sub-1% default rate. Protected ~INR 3Cr in at-risk NPA exposure that would have otherwise
resulted in write-offs on the volatile MSME loan book.
[CANDIDATE] JosephTARUN [CANDIDATE] [EMAIL REDACTED]+91 94472 3810494472 38104 [CANDIDATE]-[CANDIDATE]-pm [CANDIDATE]-[CANDIDATE]-pm
VP Product and Growth | Jai Kisan (Bharat Khata) | Dec 2022 – Feb 2024
• Built a Rural-First Credit Infrastructure: Led the 0-to-1 launch of an agri-specific Loan Origination System (LOS)
tailored for the Bharat market, digitizing and disbursing INR 500Cr+ in annualized agricultural credit. Contributed
directly to ~INR 12–15Cr in annual interest and processing fee revenue by enabling the company to scale
from pilot to full-state rollout across 3 agri-corridors.
• Engineered a Localized Document Intelligence Pipeline: Deployed spatial parsing and SLMs tuned for regional
anomalies to autonomously validate non-standard handwritten mandi receipts and vernacular trade ledgers,
collapsing loan approval TAT from 2 weeks to under 3 hours. Reduced field verification costs by INR 80L/year
while enabling 4x throughput on the same operational headcount.
• Formulated a Yield-Backed Underwriting Model: Built an ML scoring algorithm integrating crop seasonality,
regional monsoon telemetry, and real-time commodity pricing to dynamically scale dealer working capital limits,
holding portfolio delinquency below 1.2%. Prevented an estimated INR 6Cr in NPA losses over the operating
period by replacing collateral-based approvals with data-driven risk controls.
• Overcame Digital Literacy Barriers via Voice AI: Launched a multilingual conversational AI assistant (English,
Hindi, Nepali) for self-serve KYC and origination across tier-3/4 markets, slashing CAC by 35% and reducing
field-agent dependency. Saved ~INR 1.5Cr annually in field-agent and acquisition costs, directly improving
unit economics per loan.
• Orchestrated End-to-End SCF Digitization: Deployed a fully closed-loop Supply Chain Finance engine integrated
with anchor corporate ERPs for instant EIPP and automated reconciliation, processing INR 1,000Cr+ in
annualized invoice volume. Unlocked ~INR 8Cr in annual SCF fee revenue (0.8% platform fee on facilitated
invoice financing) from anchor-side monetization.
VP Product and Marketing | Payme | Feb 2022 – Dec 2022
• Spearheaded a High-Velocity Consumer Credit Funnel: Integrated multi-nodal API gateways (CIBIL, NSDL,
Account Aggregators) for sub-2-minute algorithmic approvals, processing 50,000+ monthly micro-loan
applications. Supported ~INR 10Cr in monthly disbursals, generating ~INR 1.2Cr/month in processing and
interest fee revenue at the platform's peak run-rate.
• Defeated Fraud with Computer Vision: Architected a deep learning OCR pipeline to parse, flatten, and
cross-verify PAN cards and bank statements, plummeting identity fraud by 40% and pushing extraction accuracy
to 95%. Prevented ~INR 2.5Cr in estimated annual fraud-related write-offs while reducing manual KYC ops
costs by INR 40L/year.
• Architected an Alternative ML Credit Engine: Deployed predictive ML models analyzing transactional SMS
patterns, utility payments, and device telemetry to safely approve 30% more credit-invisible applicants while
holding NPAs below 2.5%. Grew the addressable borrower base by 30%, generating ~INR 3Cr in incremental
annual revenue from applicants who would have been declined under traditional bureau scoring.
• Optimized Mobile Conversions & Digital Collections with NLP: Deployed an omnichannel NLP chatbot for
frictionless Video-KYC completion and automated WhatsApp collection nudges, boosting funnel conversion by
22% and reducing 30-day delinquency by 15%. Translated to ~INR 1.8Cr in recovered collections on an
annualized basis and a measurable reduction in the cost-to-collect per delinquent account.
Product Head (Co-Founder) | Bonoape | Jan 2019 – Feb 2022
• Architected Retail Middleware: Founded and scaled a B2B2C FMCG platform; deployed a free mobile-first smart
billing app (React Native) that invisibly digitized inventory for 400+ daily active Kirana stores, with a 400% usage
spike during early-pandemic supply chain disruptions.
• Engineered a High-Speed OCR Pipeline: Deployed an asynchronous computer vision engine (Tesseract, AWS
Lambda) custom-trained on degraded thermal paper to parse 30-item distributor invoices in under 15 seconds.
Eliminated ~INR XXL/year in manual data-entry costs for distributor clients and became a key stickiness driver
for paid tier retention.
• Planned a Multi-Tiered Monetization Engine: Built a scalable B2B2C revenue model featuring premium retailer
SaaS, an algorithmic FMCG ad-bidding network, and a high-concurrency inventory API (Node.js/Redis) for
hyperlocal fleet dispatch.
• Delivered Real-Time Analytics: Launched an enterprise dashboard granting FMCG giants (Unilever, MTR)
visibility into 500+ micro-markets. Scaled to Pre-Series A term-sheet discussions before executing a structured,
ethical shutdown , migrating users and returning angel capital following the pandemic-related loss of a key
co-founder.
Product Owner (AI Implementation) | Anakage (Accenture APAC) | Feb 2017 – Oct 2018
• Built a 'Shift-Left' IT Automation Engine: Spearheaded an intelligent endpoint management platform integrating
with ServiceNow and Jira to autonomously deflect 40% of routine IT requests for Fortune 500 clients. Contributed
to ~$2M+ in annual client helpdesk cost savings per enterprise account, directly supporting renewal and
upsell cycles and a 25% increase in contract renewal rates. Launched an Enterprise Software Telemetry
Dashboard:Directly drove a 25% increase in enterprise contract renewals by proving productivity gains,
contributing to a multi-million dollar expansion in Accenture APAC's managed services book.
• Deployed a 'Self-Healing' RPA Infrastructure: Built a background telemetry and RPA pipeline that autonomously
resolved endpoint compliance gaps, VPN drops, and cache bloat, slashing MTTR from 4 hours to a 1-click
10-second fix. Reduced L1/L2 support ticket volume by 40%, saving clients ~$800K/year per 10,000-seat
enterprise deployment, a key metric cited in Fortune 500 case studies.

Redacted: name ×6, email ×1, profile url ×1

Interview brief

Briefs are written when you shortlist a candidate.

Email draft

An offer or rejection draft is prepared once you make a decision. Nothing is ever sent without your approval.

Audit trail (3)

9/29/2026, 11:35:43 AMsystemanalysis completed{"PM_score":32,"SPM_score":45}
9/29/2026, 11:35:43 AMsystemrole assigned best fit{"role":"SPM","PM_score":32,"SPM_score":45}
9/29/2026, 11:29:29 AMsystemcv ingested{"file":"11_tarun_joseph.pdf","role":"AUTO","redactions":{"name":6,"email":1,"profile_url":1},"name_detected":true}