● PKCP / Finance Technologist / Regional Banking Intelligence — Internship Scope
BankShieldRegional Audit Intelligence.
A regional banking audit and analytics framework developed around the review of branch-level banking data and operations. It integrates Core Banking System analysis, credit and advances review, NPA monitoring, transaction analysis, ledger reconciliation, KYC/AML checks, exception tracking and prudential indicators to support structured identification of risks, control gaps and unusual banking activity.
RBI / Regional Banking Analytics 186 regional banking nodes CBS + credit + macro
Regional control room / supervisory cycle ● LIVE ANALYSIS
186 branches
₹1,860 Cr advances
1,84,632 transactions
1,286 exceptions
78.2 regional risk
01 Population Branches, GL, CBS, loans, deposits.
02 Re-performance Rebuild balances, interest and ratios.
03 Control tests Rules, approvals, ageing and evidence.
04 Correlation Connect exceptions across datasets.
05 Risk scoring Exposure × severity × recurrence.
06 Traceability Evidence-led finding closure.
01
Regional Banking Surveillance Command Centre FY 2025–26 / cycle 04 /
Regional risk signal Branch-weighted exception intensity against control baseline
12 MONTHS
Regional control environment Weighted score across core audit domains
78.2 / 100
Credit underwriting HIGH
CBS / operations MEDIUM
KYC / AML HIGH
Reconciliation STABLE
Regulatory reporting WATCH
Evidence completeness 91.6%
Banking nodes 186 regional branch population
Deposits ₹1,480 Cr liability base reviewed
Advances ₹1,860 Cr credit book in scope
Transactions 1,84,632 automated population tests
Exceptions 1,286 1,247 attributable / 23 open
Material exposure ₹2.16 Cr high-risk exception universe
02
Credit, NPA & provisioning surveillance loan-level re-performance / risk migration
Credit portfolio profile Independent analytical view of the regional book
₹1,860 Cr
GNPA 3.18% +0.18 pp YoY
NNPA 1.09% +0.07 pp YoY
Provision cover 71.8% watchlist
Slippage 1.94% annualised
Loan control tests Re-performance rules applied to sanction → disbursement → monitoring
38 TESTS Drawing power / stock statement 17 breaks
Sanction vs delegated authority 9 breaks
Interest recomputation 98.7%
Collateral / documentation 23 gaps
Overdue ageing 41 accounts
Early warning signals 62 accounts
Related-party linkage 8 flagged
Re-performance: expected interest, ageing, DP and approval thresholds are independently rebuilt before an exception is accepted.
03
Branch risk & Core Banking System surveillance branch → cluster → regional view
Highest-risk branches Composite ranking from credit, operations, KYC and reconciliation signals
TOP 06 Branch Risk Advances Exceptions Primary signal
BR-042 / Central 92.4 ₹386 Cr 48 DP / ageing BR-087 / Industrial 89.7 ₹512 Cr 41 AML / overrides BR-019 / Market 87.2 ₹298 Cr 37 duplicates BR-063 / North 83.8 ₹441 Cr 31 recon BR-105 / SME 81.6 ₹356 Cr 29 documentation BR-031 / Retail 79.9 ₹284 Cr 25 cash / KYC
CBS process-gap monitor Controls that deserve regional escalation
LIVE TEST LOGIC 01 / DEPOSIT MOBILISATION Three recurring process patterns detected in branch-level CBS extracts; exception recurrence is concentrated in 11 nodes.
02 / MANUAL OVERRIDE 44 process exceptions correlate with 57 amount-level override signals; 12 require maker-checker evidence.
03 / DORMANT ACCOUNT 31 activations were flagged; 6 cross the materiality threshold after transaction linkage.
04 / POSTING DELAY 17 entries show unusual end-of-day timing concentration and feed into the reconciliation queue.
04
Regulatory & macro-financial monitor capital / liquidity / asset quality / macro context
Prudential indicators supervisory dashboard
BASEL III CRAR 16.4%
Tier 1 capital 13.8%
Common equity / CET1 13.0%
Provision coverage 71.8%
Gross NPA 3.18%
Net NPA 1.09%
Liquidity & funding Balance-sheet stability signals
MONITOR CASA ratio 43.2%
Credit / deposit 76.4%
High-cost deposit share 10.9%
Liquidity buffer 1.31×
Stress gap ₹142 Cr
Macro context Context layer for interpreting banking signals
6 MPC CYCLES CPI trend 5.3% → 4.7%
WPI trend 2.8% → 2.1%
Repo rate 6.50%
Yield curve +18 bp
Fiscal / borrowing watch WATCH
OMO / liquidity stance STABLE
05
Advanced audit analytics & anomaly engine numbers → relationships → clusters → review priority
Multi-dimensional exception matrix Independent tests converge across amount, timing, process, counterparty and user behaviour
5 × 5
Signal
Amount
Timing
Process
Counterparty
User
Threshold proximity
38
11
24
4
13
Duplicate / near-duplicate
118
36
79
22
17
Dormant activation
31
31
19
14
9
Manual override
57
18
44
7
12
Round-number / velocity
64
21
28
16
19
Exception intelligence Signals are promoted when independent tests converge
AUTO-RANKED 01 / CONVERGENCE 24 threshold-near transactions also show process exceptions — higher priority than isolated rule hits.
02 / CLUSTERING 118 duplicate-pattern hits collapse to 17 clusters after date, amount and counterparty normalisation.
03 / VELOCITY 19 user-linked cases show transaction bursts outside the branch operating profile.
04 / ROUND-NUMBER TEST 64 amount-pattern signals require corroboration; the test is treated as a lead, not proof of misconduct.
05 / DATA QUALITY 255 exceptions lack a complete evidence reference and remain open until source support is attached.
Analytical rules 64 continuous control tests
Population tested 1,84,632 transaction-level records
Exceptions matched 1,247 98.8% attributable
Clusters 352 multi-signal groups
High-risk cases 84 priority review queue
Evidence gaps 255 open / not auto-cleared
06
Forensic reconciliation & independent re-performance reconstruct the number / don't trust the first number
GL → CBS → sub-ledger bridge balance reconstruction
₹128.46 Cr Reported difference ₹1.86 Cr
+1.45% Unexplained residual
₹17.2 L Independent rebuild: ₹6.8L + ₹4.2L + ₹3.1L + ₹0.8L + ₹3.7L = ₹18.6L, tying to the independently rebuilt variance.
Evidence-led forensic chain Every mismatch is decomposed before a conclusion is made
TRACEABLE
01 / SOURCE GL, CBS, sub-ledger
02 / REBUILD independent total
03 / COMPARE absolute + %
04 / DECOMPOSE cause hypotheses
05 / TRACE voucher / evidence
06 / CONCLUDE finding / closure
Competing explanations / ranked
01 / CUT-OFF ₹29.4L — supported by posting timestamps and period-end feed behaviour.
02 / UNPOSTED CBS ₹18.2L — requires voucher and batch-level evidence before closure.
03 / RECON BREAK ₹13.3L — linked to unmatched source-to-ledger items.
04 / RESIDUAL ₹4.9L — remains OPEN pending primary evidence.
07
KYC, AML & transaction surveillance customer → transaction → branch → network
Customer / transaction control tests Rule-based surveillance with evidence escalation
52 TESTS High-risk KYC review overdue 37
Structuring / threshold proximity 38
Dormant account activation 31
Rapid in / rapid out movement 26
Duplicate beneficiary / counterparty 17
Unusual weekend / after-hours activity 21
Maker-checker override 12
Transaction graph Relationship-led review rather than isolated alerts
NETWORK VIEW Cluster A: 7 accounts → 3 counterparties → 2 branches → 14 linked transactions → ₹24.8 Cr gross movement.
Cluster B: 5 users → 2 branches → 11 overrides → 17 near-duplicate settlements.
Cluster C: 6 dormant activations → 4 counterparties → ₹2.16 Cr movement → 6 priority cases.
Rebuild network Refresh AML queue
08
Suspense, reconciliation & exception ageing break → age → assign → resolve
Suspense account ageing Long-pending items are escalated by value and age
₹36L
Exception ageing / closure discipline Open items remain visible until evidence-backed closure
1,142 OPEN 0–30 days 652
31–60 days 247
61–90 days 196
91–180 days 96
180+ days 52
Overdue / escalated 214
09
Risk-based sampling & audit planning population → materiality → sample → evidence
Sampling engine Risk-weighted selection rather than arbitrary samples
Total transaction population 1,84,632
High-risk population 8,462
Selected substantive sample 248
Selected high-risk sample 86
Branch coverage 186 / 186
Override / judgement sample 74
Selection logic Exposure and risk drive review priority
RISK WEIGHTED HIGH EXPOSURE Large-value, threshold-near or material balance movements receive higher selection weight.
CONTROL FAILURE Repeated maker-checker, documentation or reconciliation breaks increase sample priority.
RECURRENCE Prior-cycle exceptions and unresolved findings receive targeted re-testing.
10
Exception register & management action risk × exposure × ageing × evidence × owner
ID Area Observed condition Exposure Age Owner Evidence Risk
CR-0482 Credit Drawing power not revised after latest stock statement ₹4.28 Cr 42d Credit Stock + DP calc HIGH AML-0194 KYC / AML High-risk customer review overdue ₹2.16 Cr 37d Compliance KYC trail HIGH REC-0831 Recon Debit item outside ageing threshold ₹48L 96d Operations Advice / GL MEDIUM SUS-0312 Suspense Long-pending entry lacks supporting narration ₹86 L 181d Branch Voucher pending HIGH TX-0227 Payments Near-duplicate settlement with same counterparty ₹36L 12d Payments Payment + reversal HIGH CBS-0171 CBS End-of-day posting delay creates reconciliation mismatch ₹28L 28d IT / Ops Batch log MEDIUM
Generate exception pack Refresh action queue
11
Evidence trail & reproducible audit log source → test → exception → evidence → conclusion
14:32:08 INGEST-01 GL, CBS, loan, deposit and transaction extracts aligned to branch master [OK]
14:32:41 TEST-17 64 analytical control rules executed against 1,84,632 records [OK]
14:33:06 MATCH-09 1,247 exceptions linked to primary cause, source record or evidence reference [OK]
14:33:29 RISK-04 84 cases promoted to high-risk queue using exposure, severity and recurrence [OK]
14:34:12 VAR-04 ₹36L GL/CBS difference independently reconstructed and decomposed [REVIEW]
1434:36 TRACE-22 ₹17.2 L residual remains unexplained — evidence request generated [OPEN]
14:35:04 REPORT-08 Branch heatmap, exception register and management action pack prepared [READY]
Run full control scan Generate workpapers Refresh evidence queue Export risk snapshot
12
Portfolio narrative RBI / Regional Banking Analytics → BankShield design
Regional analysis What the interface demonstrates
Branch-level population analysis: compare banking nodes, identify concentration, isolate variance and escalate unusual patterns instead of reviewing every record equally.
CBS + process intelligence From raw system data to control gaps
Core Banking System lens: test posting behaviour, deposit mobilisation, maker-checker controls, overrides, reconciliation breaks and evidence completeness.
Macro + prudential context Numbers need context
Supervisory-style context: connect credit growth, asset quality, capital ratios, liquidity, CPI/WPI, rates and funding signals before interpreting a branch-level exception.