Languages
- Go
- Java
- TypeScript
- JavaScript
- Python
- SQL
Founding Engineer · Hyderabad, India
SharatChandra
Distributed systems for money. Applied AI for work that still needs judgment.
I build distributed financial infrastructure that has processed ₹30B+, developer platforms that compress delivery, and agentic systems whose claims can be tested.
Skills
The stack, without the keyword cloud.
Tools grouped by the work they help me do: move money safely, operate distributed systems, ship quickly, and make AI behavior measurable.
Backend & systems
- Spring Boot
- Node.js
- NestJS
- FastAPI
- REST APIs
- gRPC
- Microservices
- System design
Data & streaming
- MongoDB
- PostgreSQL
- Redis
- Kafka
- Redis Streams
- Amazon Kinesis
- Apache Iceberg
- S3
- OLAP
Cloud & delivery
- AWS
- Lambda
- ECS
- Docker
- Kubernetes
- CI/CD
- Bitbucket
- Observability
- Fault tolerance
AI engineering
- LLM integration
- Voice agents
- Streaming ASR/TTS
- MCP
- Tool calling
- RAG
- Guardrails
- Evaluation
Foundations
- Distributed systems
- Concurrency
- asyncio
- API design
- Design patterns
- Automated testing
- Idempotency
Experience
Systems with measurable consequences.
Scope, implementation, and outcomes. No abstract responsibility lists.
01 · DEC 2025 · PRESENT
ASBL
Founding Engineer, SDE II · Hyderabad
Own the finance platform and applied AI systems across architecture, delivery, and operations.
₹30B+payment volume
95%org transactions captured
<100 msanalytics response
- Integrated HDFC Bank APIs into idempotent payment and bank-reconciliation workflows processing ₹30B+ in payment volume.
- Launched ASBL's Financial OS for invoicing, reconciliation, budgeting, loan tracking, and prepayments - capturing 95% of organisation-wide transactions and preventing up to ₹100M per month in revenue leakage.
- Engineered a shared Go communications platform on Redis Streams sustaining 1K TPS and 500K-recipient campaign bursts with Kubernetes autoscaling and vendor failover.
- Built a Linear-based developer platform that automated CI/CD, ephemeral environments, velocity tracking, and releases, cutting lead time by 96% from 48 hours to 2.
GoNestJSMongoDBRedis StreamsKubernetesAWSFastAPIMCP
Applied R&DDuet / AiraApplied voice-AI R&D for an ASBL Broadway concierge: an interruptible ASR → LLM → TTS system with typed cancellation, deterministic safety gates, and 148 focused tests. An engineering prototype, not an official production product.
02 · MAY 2023 · DEC 2025
Inncircles
Tech Lead, SDE II · Hyderabad
Scaled a backend, its operating model, and the team responsible for it.
100K+requests each day
1 → 26team growth
30%lower p99
- Architected Kafka and gRPC services for payments, invoicing, cash flow, and workflows at 100K+ requests a day, reducing p99 by 30%.
- Led the monolith-to-microservices migration on AWS ECS with Docker and a service mesh, making deployments 60% faster.
- Built core property-management modules in Java, Spring, and Node; grew the team from 1 to 26, led 12, and earned two promotions in under two years.
JavaSpring BootNode.jsKafkagRPCPostgreSQLDockerAWS ECS
Selected projects
Four systems, each with a clear boundary.
I love hackathons because they remove the distance between an idea and a working demo. These projects keep that bias toward building, with clearer evidence and harder boundaries.
01 · ASBL applied R&D · Open sourceengineering prototype
Duet / Aira
An interruptible, guarded voice system for a real-estate concierge.
148 focused tests · 17-scenario golden set · 198 ms cancellation in the latest synthetic smoke
Python 3.12asyncioSarvam ASR/TTSGeminiLangfusePrometheusPostgreSQL
02 · Architecture and implementationworking system
Hermes Brain
A Git-backed learning and memory harness shared across five AI clients.
5 AI clients · Source provenance · Human approval gates · Regression evaluation
GitMCPMarkdownJSONLSchema validationEvaluation gates
03 · Product, architecture, and implementationBuild Week prototype
LazyTax
Evidence-first Indian tax casework for taxpayers and practising CAs.
9 local MCP tools · 11/11 Trust Lab controls · Source-indexed Tax Proof Pack
TypeScriptNode.jsMCPZodPlaywrightDeterministic rules
04 · Product, systems, and implementationworking prototype
Stormwise
A personal flood digital twin for Bengaluru households and commutes.
Live public data · Route-level risk · Desktop, mobile, touch-target, and WCAG coverage
Next.jsTypeScriptMapLibreZodOSRMOpen-MeteoGloFASGemini
Evidence
A few external signals.
2nd / 400Google PromptWars 2026
WinnerSmart India Hackathon 2022
3Scopus indexed research papers
Technical notes
What I am learning in public.
Concise notes backed by sources, implementation work, or both.
Benchmarks that can hurt youI published the table where my own system loses. Here's why that was the whole point.2026The reconciliation mindsetMatch 99%, surface the 1%. The useful pattern hiding inside boring finance work.2026Two days to two hoursWhat actually made releases 24× faster. It wasn't the pipeline.2026
Browse the learning archive →Contact
Building a system where reliability and product judgment both matter?
I am open to thoughtful conversations about founding engineering, platform work, financial systems, and applied AI.