Mudassir Ahmed

Senior Backend Engineer — Core Banking Systems
Hyderabad, India · mudassirahmed03@gmail.com · LinkedIn

About

I build core banking and fintech systems — distributed ledgers, payments, compliance, and high-throughput transaction engines. Four-plus years at Atlas / HugoHub, a Banking-as-a-Service platform, working end-to-end from new service design through production hardening and performance work. I care about throughput, latency, correctness, and reliability — the things that decide whether financial infrastructure survives contact with real money.

By the numbers

25 → 1,500+ TPS
Ledger throughput on a single instance — 60× improvement via parallelism and execution isolation
~30–50%
Inter-service latency reduction from gRPC + Java Virtual Threads migration
~8 ms
Envelope-encryption (KEK/DEK, AES) latency for PII at rest
< 30 ms
Product Factory parameter evaluation across fees, taxes, limits
6 services
Built from scratch: Accounts, Hugosave App, Compliance, Cash, Audit, Billing
4,500+ commits
Sustained ownership across 21 backend domains since 2022 — Cash, Ledger, Compliance, Audit, Billing, KMS, Risk, Profile, and more

Experience

Senior Backend Engineer — HugoHub (BaaS Platform)
Jan 2024 – Present
Atlas · Hyderabad · Distributed Systems, Ledger, Security
  • Ledger — auto-scaling rewrite. Re-architected the ledger module around bucket-partitioned balance tracking with automatic expansion/compaction, contention metrics, async snapshotting with global ordering guarantees, and cross-pod expansion cooldowns. Took single-instance throughput from ~25 → 1,500+ TPS. Shipped admin repair tooling for snapshot reset and bucket reconcile.
  • gRPC + Virtual Threads migration. Drove the platform-wide move from Lognet to Spring gRPC and rolled out Java Virtual Threads across 11 services — ~30–50% latency reduction under load and 5–10× concurrency for IO-heavy paths. Wrote MDCAwareVirtualThreadExecutor to keep tracing context intact across virtual-thread boundaries.
  • Cash Workflow Engine. Refactored cash account creation and transaction initiation into a parallel step-group engine with sub-workflows for ledger writes, balance validation, and limit checks. Eliminated bespoke orchestration code across the Cash domain.
  • Product Factory. Configuration-driven engine for fees, taxes, charges, limits, yields with sub-30 ms parameter evaluation. Cut new-product launch time by ~60–70%. Foundation reused across Cash, Card, CMS, and Investment.
  • Encryption platform. Centralized envelope encryption (KEK/DEK, AES) for PII at rest at ~8 ms overhead, with proto-value encryption integrated into the DAO layer.
  • Billing service. Built end-to-end metering and invoicing from scratch — @MeteredBilling aspect for event capture, 12+ CRUD domains (BillingConfig, Tier, RuleGroup, CostConfig, Invoice, CustomerBalance), pricing pipeline, invoice engine with line-item breakdowns, billing-cycle worker, gRPC service. ~7,400 lines in the foundational commit.
  • Audit platform. Stood up MySQL-backed audit logging across all services with Kafka transport, Loki integration, and dedicated ag-audit / dt-audit consumer microservices. Centralized into common config so every service inherits it automatically.
  • Reliability hardening. Replaced cache-based mutex with Redisson distributed locks; enforced idempotency at the DB layer via MySQL triggers for accounts and transactions; routed ad-hoc queries to writer to avoid replication lag; made @Transactional optional for single-DB-op paths to cut transaction overhead.
  • Engineering leadership. Onboarded and mentored engineers across Cash, Compliance, and core platform; led the 80% test-coverage push across Risk, Compliance, Ledger, Investment, and Reward.
Backend Engineer — Hugosave (Consumer Product)
Jan 2022 – Dec 2023
Atlas · Hyderabad · Product, Migration, Compliance
  • Monolith → microservices. Re-architected the banking backend from Python/AWS Lambda to Java/Spring Boot on EC2. Eliminated cold starts, enabled state-driven workflows — 2–3× throughput.
  • Service bootstrap. Built Accounts, Hugosave App, and Compliance from zero — KYC integration, account onboarding, signups, pots, referrals, full DAO/controller stacks. ~146 files / 7,375 lines in the foundational commit alone.
  • Cash service. Built the Cash domain end-to-end from scratch: accounts, beneficiaries, transactions with full lifecycle (initial, refund, multi-deposit, pay-beneficiary), settlements, overdrafts, multi-currency, transaction limits. ~5,900 lines of BDD test coverage. Zero-to-prod with migration controls.
  • Distributed state machine. Designed the lifecycle framework that backs every backend workflow — improved reliability and dev velocity by ~30–40%.
  • Intent processing engine. Step-based async processor framework with mutex-based concurrency control, intent references, types, and views — used for every banking workflow (account opening, transactions, cards, investments, roundups, live user migration).
  • QLDB → MySQL migration. Designed MysqlProvider and the DbApi abstraction; migrated all domain models (intents, schedules, transactions, pots, users) from QLDB-specific to generic. Architectural pivot for the entire platform.
  • Live platform migration. Built the online migration pipeline that moved active users to the next-generation platform without downtime — vault opening, accounts-to-trust for premium users, schedule/beneficiary migration, compliance close/resubmission, retry logic for blocked users and edge cases.
  • Compliance & risk engine. Pre-transaction rule engine at ~100 ms with automated dispute workflows; TruNarrative KYC integration with Singapore document validation (NRIC/FIN); Trust compliance flow end-to-end.
  • Platform abstractions. Built PubSubProvider (SNS), StorageProvider (S3), and the MDCAware thread-pool / fork-join executors that propagate trace IDs across async paths — adopted org-wide.
  • Infrastructure migrations. QLDB → MySQL, then SQS → Kafka → NATS JetStream as the platform scaled.

What I work on

Architectures designed

  • Ledger Auto-Scaler — bucket-partitioned, self-balancing
  • Workflow Engine — parallel step groups, sub-workflows
  • Intent Processing Engine — step-based async framework
  • Product Factory — config-driven financial products
  • Distributed State Machine — workflow lifecycle
  • MysqlProvider — DB abstraction for QLDB → MySQL

Migrations led

  • QLDB → MySQL 2022
  • Python/Lambda → Java/Spring Boot 2022–23
  • SQS → Kafka 2025
  • Kafka → NATS JetStream 2025
  • Lognet → Spring gRPC 2026
  • Ledger architecture rewrite 2026

Stack

Languages & data
Java SQL MySQL DynamoDB Redis / Redisson Protobuf
Frameworks & RPC
Spring Boot gRPC Hibernate / JPA Virtual Threads
Messaging & async
Kafka (Strimzi) NATS JetStream AWS SQS / SNS
Architecture
Microservices Event-Driven State Machines Distributed Systems Idempotency Distributed Locking
Cloud & ops
AWS (EC2, Lambda) Docker Grafana Loki Kibana Distributed Tracing

Education

M.Tech. — Parallel and Distributed Systems
Osmania University, Hyderabad
2020 – 2022
B.Tech. — Computer Science & Engineering
Malla Reddy Engineering College, Hyderabad
2016 – 2020