Dynamic contributor with extensive software industry experience in designing
and delivering scalable end-to-end systems, microservices architectures,
and cloud-native platforms.
Own end-to-end development of a multi-tenant shipping platform and led
the monolith-to-microservices migration on Java/Spring. Cut average API
latency from 1.2s to 450ms with Redis + async processing, re-engineered
order tracking on Elasticsearch at 10K+ QPS, and integrated Moneris,
Authnet, and PayPal with a 99.95% payment success rate. Right-sized AWS
instances to unlock auto-scaling and drove team-wide AI dev-tool
adoption for a ~25% sprint velocity lift.
Delivered full-stack work across 6+ client projects using Java, Spring,
and Angular — optimizing queries and schema to enable a tier-down on
hosting without performance impact. Built a reusable UI component
library that cut front-end effort 25% across subsequent projects, and
deployed zero-downtime AWS infrastructure (EC2, S3, RDS) with
auto-scaling groups across multiple client applications.
Lambton College × TELUS
Capstone Project
Sep 2017 — Jan 2019
3 stakeholder orgs1 working prototype
Prototyped a technically viable solution addressing TELUS's core
business requirements, applying software design, database management,
and OOP coursework. Coordinated across Lambton, TELUS, and the internal
team to run kick-offs, track progress, and deliver the final prototype
demo with technical documentation and tradeoff summaries.
Pixels Galaxy
Software Engineer
Feb 2015 — Jul 2017
100K+ end users200+ tests written150+ issues resolved40% QA effort cut
Built and maintained Java/Spring backend modules for e-commerce
applications, writing JUnit and integration suites that shortened
release cycles by a full week per sprint. Ran root-cause analysis on
production issues via JIRA, automated historical billing backfills to
eliminate migration work on new projects, and authored API references
for 10+ backend modules to speed developer handoff.
Infosys
Systems Engineer
Jun 2014 — Jan 2015
4 mainframe techs
Completed specialized training in Mainframes — COBOL, JCL, CICS, DB2 —
then performed enhancements and internal support on legacy tooling.
Contributed to work-product reviews and brainstorming sessions focused
on reducing delivery defects.
Awards & Recognition
2023MVP of the Year award — eShipper
2018Placed in Top 20 of 800 in CSAT graduating cohort
2016Star of the Quarter award — Pixels Galaxy
Education
2019
Computer Software & Database Development (CSAT)Lambton College, Ontario, Canada
2014
Bachelor of Technology, Information TechnologyGuru Gobind Singh Indraprastha University, New Delhi, India
About this background
Live Agent Network
The background depicts a multi-agent AI system — the architectural
pattern behind modern agentic applications. Each labeled node is a distinct
agent or component, and the lines between them represent active communication
channels. The glowing pulses traveling along those lines are live messages
passing between agents in real time.
This is how production AI systems are actually built today. Rather than one
monolithic model doing everything, tasks are broken down and routed to
specialized agents that each handle a specific job:
Orchestrator — coordinates the overall workflow, deciding which agents to invoke and in what order
Planner / Router — breaks down complex goals into subtasks and routes them to the right agent
Retriever / Reranker / Embedder — find and surface relevant information from vector stores (the RAG pipeline)
Reasoner / Critic / Validator — evaluate outputs, check for errors, and refine results before returning them
Tool Agent / Executor / MCP — reach outside the model to call APIs, run code, or interact with external systems
Memory / Context / Cache — maintain state across turns so the system remembers what happened earlier
Monitor / LLM — observe the system's health and provide the core language model intelligence
The connections between nodes shift as agents drift closer or further apart —
reflecting how in a real agentic system, which agents communicate depends on
the task at hand. Not every agent talks to every other agent all the time.