Scalable.
Reliable.
Infrastructure.

Systems and infrastructure engineer focused on distributed systems, resilience, and scale.

View Work

Core Technologies

Backend Engineering

Production-grade service design across Python, Go, and Node - GraphQL APIs, async task pipelines with Celery, and systems architected from the ground up for sub-120ms latency, five-nines reliability, and zero-downtime operations.

Cloud Native

Full-stack cloud ownership across AWS - containerized workloads on ECS and Kubernetes, event-driven pipelines on Kafka, infrastructure codified in Terraform, and CI/CD via GitHub Actions. Built for teams that ship fast and sleep well.

Data Systems

Data layer design that doesn't become a bottleneck - relational integrity in Postgres, sub-second analytics at scale in Apache Pinot, Redis caching, DynamoDB for high-throughput key-value, and Elasticsearch for full-text. The right store for the right workload.

Distributed Systems

Systems that hold up under pressure - event-driven architectures, reliable message queues, backpressure handling, and fault isolation patterns that keep failures local and keep the system running.

ML Infrastructure

End-to-end ML infrastructure built for production - training pipelines with live checkpoint-restore (CRIU/CUDA) for zero-loss preemption, ONNX-optimized model serving, experiment tracking, and vector search powering semantic retrieval at scale.

Observability

Full-stack observability that surfaces signal in the noise - structured logging, custom metrics, distributed tracing with Sentry and Datadog, and alert design that pages on what matters and nothing else.

Work Experience

LinkedIn

Systems & Infrastructure Engineer Intern
May 2026 – Aug 2026
  • Delivered a live ML workload checkpoint-restore capability that eliminated training loss on preemption - capturing and migrating full CPU/GPU execution state across heterogeneous machines (CRIU, CUDA, Kubernetes), directly reducing compute waste and unblocking fault-tolerant, preemptible training at scale.

Uniqode

Senior Software Engineer Aug 2024 – Jun 2025
  • Owned the full backend surface of a $2M ARR platform - driving 99.99% uptime across a distributed monolith and microservices fleet, scaling notification infrastructure to 1M+ daily cross-platform events, and reclaiming $12K/year in observability spend through targeted log infrastructure overhaul.
Software Engineer Aug 2022 – Jul 2024
  • Drove a step-change in platform performance - a ground-up analytics migration to Apache Pinot collapsed P95 query latency by 97.6%, API optimization pushed P95 under 120ms, enterprise integrations with Azure AD and Salesforce unlocked $300K+ in ARR, and built out the team through 70+ interviews and direct mentorship of 3 engineers.
Software Engineering Intern Feb 2022 – Jul 2022
  • Shipped core API and bulk operation infrastructure that cut processing latency by 40% and bulk action time by 12x through async pipelines and query re-architecture; also stood up threat intelligence coverage for Tor, VPN, and bot traffic with near-zero performance overhead.

LetsEndorse Development

Summer Intern
May 2021 – Jul 2021
  • Redesigned the KPI reporting surface with reusable React.js dashboards that compressed insight cycle time from 90 minutes to 5, and shipped 15+ React Native screens that doubled operational efficiency for delivery partners.

Selected Work

Sentinel

Distributed, Serverless Email Platform

A production email platform handling 200K+ messages/month for under $50 - a cost structure that would be impossible on traditional infrastructure. Async job queues, real-time event ingestion, multi-region replication, and a retry/DLQ architecture that treats failure as a first-class concern.

Serverless Event Driven AWS Queues

Build-flask-app

Open-Source Flask CLI (42K+ Downloads)

Eliminated the hours-long boilerplate tax on Flask projects - an open-source CLI now adopted by 42K+ developers that generates production-ready scaffolding with WebSockets, database integration, and test infrastructure wired up from day one.

Flask CLI Open Source PyPI

north

Multi-Agent LLM Orchestration System

An orchestration layer for LLM agents that treats inference cost and task continuity as core engineering concerns - dynamic model tiering routes work to the right model at the right cost, an append-only audit trail enables partial task recovery without redundant inference, and a preference engine that sharpens its routing decisions based on observed user judgment over time.

LLM Multi-Agent Orchestration AI

CuriosityAI

CuriosityAI @CalHacks 12.0

Built at CalHacks 12.0 - a semantic research discovery engine that surfaces connections humans would miss. High-dimensional embeddings in ChromaDB map the research space, dimensionality reduction collapses it into navigable 3D clusters, and LLMs generate novel, feasible research directions at the boundaries of existing knowledge.

GenAI ChromaDB Vector Search 3D Clustering

FightFlow

ML Punch Classification System

A rigorous ML study showing the right inductive bias beats more compute - five architectures from scratch CNN to fine-tuned ResNet-18, all outperformed by a pose-based LSTM that leverages skeleton features invariant to lighting, background, and camera angle. A concrete demonstration that feature engineering strategy matters more than model complexity.

PyTorch MediaPipe LSTM Computer Vision

Sentry In Development

Real-Time Threat Detection Platform

A cloud-native threat detection platform built for operational rigor from day one - ONNX-optimized ML models served via Flask on ECS Fargate, React frontend on S3, infrastructure fully codified in Terraform with environment-isolated IAM, end-to-end CloudWatch observability, and automated pipelines that promote to staging and production without manual intervention.

ML Inference AWS ECS Terraform ONNX

Prism

Custom Programming Language

A programming language designed from first principles for a specific class of problem - declarative deconstruction of complex data. Built in Haskell with a Parsec grammar, pattern matching with variable binding, recursive scoping, and monadic error propagation. A demonstration of language theory applied: the right semantics make the hard problems tractable.

Haskell Language Design Parsec Interpreter

Education

Master of Science in Computer Science

San Jose State University  ·  GPA 4.0 / 4.0

Aug 2025 – May 2027

Database Systems Principles · Distributed Systems · Cloud Computing · Advanced Parallel Processing · Machine Learning

Bachelor of Technology in CS&E

Kurukshetra University

Aug 2018 – Jul 2022

Data Structures & Algorithms · Advanced Algorithms · DBMS · OOP · Computer Networks · Operating Systems

Get in Touch

Contact Info