// hi, my name is
Mahesh Shelke.
I embed, I understand, I build.
A software architect whose career has spanned genomics, e-commerce, fintech, IoT, and smart cities — every engagement a new domain, a new stack, a new problem solved from scratch. That cross-domain depth is now my biggest asset as I grow into a Forward Deployed AI Engineer — embedding with teams, architecting solutions at the frontier, and turning cutting-edge AI tools into real, working leverage in production environments.
01. About me
A personal AI assistant (Openclaw) hosted on Raspberry Pi — managing my calendar, capturing contacts at networking events, and surfacing notes on demand. Plus agentic engineering workflows using OpenAI, Claude Code, GitHub Copilot, and MCP-connected toolchains.
15+ years. Seven domains. Every major stack. I'm a language-agnostic, domain-agnostic problem solver — the kind of engineer who walks into unfamiliar territory, learns fast, and ships something that works.
Precision medicine. Fintech. IoT. Smart cities. Enterprise deep tech. Platform engineering. AI tooling. Each domain taught me something different, and the breadth of those experiences is what lets me approach new problems from angles others miss — combining patterns from genomics data pipelines with platform infrastructure thinking, or applying fintech security instincts to developer toolchain design.
Right now I'm all-in on building with AI — not just using AI tools, but engineering systems around them.
- Java · Python · JS / TS · Groovy
- Spring Boot · Node · React · Grails
- OpenAI · Claude Code · Copilot
- Openclaw · MCP · Raspberry Pi
- AWS · Docker · Kubernetes
- MongoDB · Postgres · Elasticsearch
02. Experience
Where I've worked
Senior Engineer & Developer Platform Lead @ EXL Service
2023 — PRESENT · Platform Engineering · AI Tooling
- Lead the Developer Platform team — owning Toggles (feature flags), Config Server, and NRD Server (log relay to Splunk) — plus an org-wide Spring Boot framework augmentation used across all engineering teams.
- Pioneered AI tool adoption early and independently — building autonomous incident resolution pipelines, reusable Skills, and a five-tool MCP ecosystem (SignalFX, PagerDuty, Jira, GitHub, Splunk) before it became a mandate.
- Used Claude Code, OpenAI, GitHub Copilot, and Openclaw in production workflows — presenting AI productivity findings to senior leadership and making the case for AI-native engineering as a strategic capability.
- Operated as a cross-functional problem solver across identity, SBS, and automation testing teams, bridging platform decisions with business delivery at scale.
Technical Lead (Consultant) @ Nimble Pvt. Ltd.
APR 2023 · Management Information Systems · NGO Tech
- Owned the full product lifecycle of Better Connect — a task assignment and field reporting platform for an NGO running farmer skills development programmes.
- Designed a multi-layered backend following SOA, DDD, and TDD principles; architected the Android app using MVVM — learning and applying Android architecture patterns specifically for this engagement.
- Led requirement gathering, documented functional and non-functional specs, and built out the team from scratch — demonstrating end-to-end technical leadership on a greenfield product.
Senior Application Developer @ PierianDx India Pvt. Ltd.
DEC 2019 — MAR 2022 · Precision Medicine · Genomics
- Built web services and UI components for Clinical Genomics Workspace — a next-gen sequencing reporting platform for cancer and complex inherited diseases, used by clinical teams for genomic insights and patient care decisions.
- Developed automation scripts for server-to-server database dumps with data de-identification — a critical compliance requirement in the clinical genomics domain.
- Evaluated and adopted new technologies to address evolving clinical and regulatory needs, including Elasticsearch and Logstash for high-volume genomic data search and log management.
- Produced design documentation, work breakdown structures, and effort estimations — operating as a senior technical contributor across a domain that required fast, deep learning of genomics workflows.
Tech Lead @ Future Technologies India
FEB 2019 — AUG 2019 · IoT · Smart Cities
- Investigated and implemented FIWARE — the EU smart city IoT platform — from scratch, learning NGSI v2 protocol and the full FIWARE stack to build dashboards for smart city stakeholders across waste, traffic, and utilities.
- Developed for client NEC Technologies India: a platform ingesting data from a wide range of IoT sensors and third-party systems, visualised in real-time dashboards for city administrators.
- Built a computer vision application to detect counterfeit or missing components on circuit boards — a deep-tech side project requiring fast learning of CV tooling.
- Created Java and Python simulators to test data pipelines — demonstrating the language-agnostic, "learn what the problem needs" approach that defines how I work.
Sr. Application Developer, Tech Lead & Architect @ Extentia Information Technology
FEB 2015 — SEP 2017 · Fintech · Digital Payments · IoT
- Led a team of 4 developers delivering Citrus Mobile Wallet — a fintech product where consumers load money, save cards, and make payments to merchants through the CitrusPay aggregator. Owned architecture, stack selection, and team delivery.
- Worked across payment gateways, digital currency, digital wallets, and IoT (smart bike) — four distinct sub-domains in one engagement, each requiring independent technical ramp-up.
- Analysed technology stacks for each project, designed full-stack architecture decisions, and ensured delivery quality across Groovy/Grails backend and frontend systems.
- Awarded Rising Star of the Year 2015 and recognised for hard work and dedication in 2016 — building a reputation for shipping under pressure and taking ownership.
Senior Software Engineer @ Persistent Systems Ltd.
MAY 2011 — NOV 2014 · Enterprise Deep Tech · Mobile App Management
- Built the Persistent Enterprise App Store — a deep tech mobile app management platform covering iOS and Android enterprise app distribution, resigning, authentication, review workflows, dashboards, and dynamic reporting.
- Individually developed a Grails web application for iOS app resigning with Enterprise License — a technically complex piece requiring deep learning of Apple's provisioning and code-signing infrastructure.
- Owned backend and frontend development for Authentication/Authorization, App Review, Dashboard, and Dynamic Reporting modules — full-stack ownership across a 3.7-year tenure.
- Worked with LDAP, Apigee API Management, Jenkins CI, and Jasper Reports — building familiarity with enterprise-grade tooling across security, API governance, and reporting.
03. Skills
Tech I work with
AI & Agentic
- Claude Code (CLI)
- OpenAI API
- GitHub Copilot
- Openclaw (local LLM)
- MCP Servers
- N8N Automation
- Google AI Studio
- Skills Engineering
Languages
- Java 8+
- Python
- TypeScript / JavaScript
- Groovy
- SQL
Backend
- Spring Boot / Spring
- Node.js / Express
- Grails / Hibernate
- REST APIs / Microservices
- Sequelize / ORM
Frontend & Mobile
- React
- Android (MVVM)
- HTML / CSS / Bootstrap
- D3 / jqPlot
- jQuery / Ajax
Data & Infra
- PostgreSQL · MongoDB
- Elasticsearch · Redis
- Docker · Kubernetes
- AWS · CI/CD
- Kibana · Logstash
Domains
- Precision Medicine
- E-Commerce
- Fintech / Payments
- IoT / Smart Cities
- Enterprise Platforms
04. Projects
Things I've built
Openclaw — Personal AI Assistant
Personal · LiveAI · Python · Raspberry Pi · OpenAI · Local LLM
- Built a personal AI assistant using Openclaw (local LLM) + OpenAI, self-hosted on a Raspberry Pi — offline-capable, privacy-first, always on, zero cloud dependency.
- Captures contacts at networking events from voice notes — stores them, enriches context from conversation, and lets me search and surface people by topic or date on demand.
- Manages calendar entries, sets context-aware reminders, and takes structured notes — acting as a persistent second memory for professional life.
- Entirely self-built and self-hosted — designed the architecture, chose the stack, wired the integrations. A personal proof-of-concept for AI as ambient infrastructure.
One Command Full Resolution — Autonomous Incident Pipeline
PlatformClaude Code · MCP · SignalFX · PagerDuty · Splunk · GitHub
- A single CLI command triggers full end-to-end incident resolution — alert detection in SignalFX, PagerDuty correlation, Splunk log analysis, root-cause identification, and a GitHub PR — without human intervention.
- Built on top of the MCP ecosystem to give Claude Code rich cross-tool awareness in a single session, enabling it to reason across the complete incident lifecycle.
- Reduced mean-time-to-resolution for routine platform incidents significantly — transforming what used to be a multi-person, multi-tab toil exercise into a single autonomous command.
- Showcased as a key AI adoption outcome to senior engineering leadership as evidence of agentic AI delivering measurable operational leverage.
MCP Ecosystem — Five-Tool AI Context Layer
ActiveMCP Protocol · Claude Code · SignalFX · PagerDuty · Jira · GitHub · Splunk
- Designed and wired a five-tool MCP server ecosystem for Claude Code: SignalFX for metrics, PagerDuty for incidents, Jira for tickets, GitHub for code context, and Splunk for logs.
- Solved the core problem of fragmented tool context — an engineer (or AI agent) previously had to context-switch across five browser tabs to diagnose an issue; now all context flows into a single agentic session.
- Foundation layer for the One Command Full Resolution pipeline — the MCP ecosystem is what makes cross-tool autonomous reasoning possible.
- Documented and shared as a reusable pattern for other teams looking to build context-rich AI workflows on top of existing enterprise toolchains.
Toggles — Feature Flag Server
ProductionSpring Boot · Java · REST API · AWS
- Production feature flag infrastructure enabling runtime configuration of feature states across services without redeployments — supports gradual rollouts, kill-switches, and environment-scoped flags.
- Integrated directly into the org-wide Spring Boot framework augmentation, giving every team zero-friction access to feature flag capabilities with no additional wiring.
- Part of a three-service Developer Platform stack — designed for reliability and minimal operational overhead so the team isn't on-call for the infrastructure that manages on-call response.
NRD Server — Log Relay to Splunk
ProductionSpring Boot · Java · Splunk HEC · Observability
- High-throughput log relay service that aggregates, enriches, and forwards structured application logs from multiple services into Splunk for centralised observability.
- Designed for reliability and low-latency forwarding — handles bursty log volumes without dropping events, with graceful backpressure handling.
- Third pillar of the Developer Platform infrastructure stack alongside Toggles and Config Server — giving the org a consistent, owned observability pipeline rather than relying on ad-hoc log shipping.
Clinical Genomics Workspace — Precision Medicine Platform
Healthcare · GenomicsSpring Boot · Grails · Elasticsearch · MongoDB · Logstash · AWS
- Reporting platform for next-generation sequencing tests covering cancer and complex inherited diseases — used by clinical teams to deliver genomic insights for precise patient care decisions.
- Built web services and UI components for the core workspace application, and developed automation scripts for data de-identification — a critical compliance requirement in the clinical domain.
- Implemented Elasticsearch-powered search across high-volume genomic datasets — requiring fast learning of both the genomics domain and distributed search architecture.
- Domain was entirely new — picked up clinical genomics workflows, regulatory constraints, and the clinical data model from scratch and contributed meaningfully within weeks.
Citrus Mobile Wallet — Fintech Payments
FintechGroovy · Grails · Java · REST · PostgreSQL
- Full-stack fintech mobile wallet — consumers load money, save credit/debit cards, and make payments to merchants via the CitrusPay aggregator.
- Led a team of 4 developers: owned architecture design, technology stack selection, and end-to-end delivery from inception to production launch.
- Worked across payment gateways, digital currency, digital wallets, and IoT (smart bike) sub-domains simultaneously — each requiring independent technical ramp-up within the same engagement.
- Awarded Rising Star of the Year 2015 during this period — recognised for shipping under pressure and taking full ownership of outcomes.
NEC IoT Smart City Dashboard — Smart Cities
IoT · Smart CitiesFIWARE · Elasticsearch · Spring Boot · Docker · Python · Apache Nifi
- Developed for NEC Technologies India — an IoT platform ingesting data from a wide range of sensors and third-party systems via NGSI v2 protocol (FIWARE), visualised in real-time dashboards for city administrators.
- Investigated and implemented the entire FIWARE smart cities stack from scratch — learned the EU IoT platform, its components, and NGSI protocol independently before any team guidance was available.
- Built a computer vision application to detect counterfeit or missing components on circuit boards — a deep-tech side project requiring fast acquisition of CV tooling in Python.
- Developed Java and Python data pipeline simulators for testing — demonstrating the language-agnostic approach of learning whatever the problem demands.
Persistent Enterprise App Store — Deep Tech Platform
Enterprise · Deep TechJava · Grails · LDAP · Apigee · Jenkins · iOS Signing
- Enterprise mobile app management platform for iOS and Android distribution — built Authentication/Authorization, App Review, Dashboard, and Dynamic Reporting modules across a 3.7-year tenure.
- Individually developed a Grails web application for iOS enterprise app resigning with enterprise license — a technically complex piece requiring deep learning of Apple's provisioning and code-signing infrastructure.
- Worked with LDAP for directory authentication, Apigee for API management, Jenkins for CI, and Jasper Reports for dynamic reporting — broad enterprise toolchain across security, governance, and reporting.
- Full-stack ownership across backend and frontend — the breadth of this project laid the foundation for the domain-agnostic engineering approach I carry today.
05. Writing
Thoughts on engineering
- From Copilot to Claude Code — what agentic AI actually changes for a platform team
- One command, zero pages: designing the autonomous incident resolution workflow
- MCP servers in practice — connecting five tools into a coherent AI context layer
- Why every platform team should own a feature flag server (not rent one)
- Measuring developer productivity without becoming the team that counts lines of code
06. What's next?
Let's talk.
I'm always up for conversations about platform engineering, AI-native workflows, developer experience, or the next hard problem worth solving. Whether it's a collaboration, an opportunity, or just a good tech discussion — reach out.