Introduction:
Axis AMC is the asset management subsidiary of Axis Bank, India’s third-largest private sector bank. Axis AMC/Axis Mutual Fund is a premier, SEBI and AMFI-regulated financial powerhouse headquartered in Mumbai. Axis AMC’s investment philosophy is anchored in rigorous bottom-up research, quantitative risk management, and disciplined portfolio construction.
Client:
Axis Asset Management Company Ltd. (Axis Mutual Fund)
Industry:
BFSI / Capital Markets – Asset Management (Risk & Quantitative Analysis)
Offering:
24/7 Managed Services & Consulting
AWS Services:
- AWS Batch, EC2 (Hpc7g, C7gn, Spot Instances) – HPC burst compute; AWS Step Functions – Risk pipeline orchestration
- Amazon Aurora PostgreSQL, DynamoDB – Portfolio data layer; Amazon MSK (Kafka) – Real-time market data ingestion
- Amazon EKS – Containerized risk engines; FSx for Lustre – High-performance shared file system for Monte Carlo
- Amazon S3 (Object Lock), AWS KMS, Neptune, SageMaker, Kinesis, ElastiCache, Redshift, QuickSight – Model governance, encryption, surveillance, analytics
The Challenge:
SEBI stress testing mandate post IL&FS (2018) and DHFL (2019) credit crises required daily execution of 50+ liquidity stress scenarios across 35+ debt schemes – taking 11 hours on legacy infrastructure. Monte Carlo simulations for portfolio VaR (10,000+ paths) ran sequentially in 8–12-hour overnight batches, with results arriving too late for next-day fund management. Intraday factor risk attribution and real-time derivatives Greeks computation were computationally infeasible on shared on-premises hardware.
The Solution: Strategic Implementations
Deployed a SEBI-compliant HPC risk platform with elastic burst compute on AWS – scaling from near-zero to 50,000+ vCPUs in under 3 minutes. Disaggregated risk pipelines into containerized workloads on EKS with independent scaling; enabled real-time market data integration via Direct Connect from NSE/BSE. Implemented S3 Object Lock (WORM) for tamper-proof model governance and BYOK KMS encryption ensuring AWS has zero access to plaintext fund data.
Technical Impact:
- 10x speed-up in Monte Carlo runtimes (8-12 hrs → <30 min); SEBI debt stress testing reduced from 11 hours to 28 minutes
- Real-time intraday VaR updates every 15 minutes; factor risk attribution and derivatives Greeks refresh every 60 seconds
- Insider trading surveillance alert latency reduced from T+2 days to <4 hours using Neptune graph database and SageMaker GNN model
- 70% reduction in compute cost via Spot Instances; disaggregated risk pipelines eliminate silent scheduling dependencies and model staleness
The Business Impact: Impact Engineered by Rapyder
- 38% reduction in total risk infrastructure cost (₹4.2 Cr on-premise CapEx eliminated; pay-per-model-run pricing with zero idle compute cost)
- Daily SEBI-compliant stress testing now fully operationalized; SEBI audit preparation reduced from 10 weeks to 2 weeks (80% reduction)
- Fund managers now make same-session trading decisions with real-time risk updates; 3 new quantitative strategies launched (previously computationally infeasible)
- 100% SEBI Cloud Framework compliance from Day 1; tamper-proof model governance with S3 Object Lock enabling seamless regulatory audits.
