Portfolio — Doc. No. KM-2026 / Rev. A New York, NY
Available for new-grad SWE roles · 2026

KrishMakadia

MS CS @ Stony Brook · Distributed Systems & Full-Stack Engineer

I build high-performance distributed systems and polished, production-grade full-stack products.

View projectsGet in touch
makadiakrish@gmail.com GitHub LinkedIn
Distributed systems✦Raft consensus — 526K txn/s✦Full-stack engineering✦Applied AI · RAG✦1.2 TB satellite imagery / day✦Systems that survive failure✦Distributed systems✦Raft consensus — 526K txn/s✦Full-stack engineering✦Applied AI · RAG✦1.2 TB satellite imagery / day✦Systems that survive failure✦
FIG. 01About

A bit about me

I'm a Master's student in Computer Science at Stony Brook University (graduating May 2026), focused on the systems that make software fast and reliable at scale.

My research rebuilds replication for Mako, a geo-replicated key-value store — I replaced its Paxos layer with a custom C++ Raft backend and pushed it to 526K committed transactions/second. Alongside research, I've shipped production software as a full-stack engineer at Stony Brook's Web Dev department and built large-scale satellite-imagery ML pipelines at ISRO.

I like working across the stack: from consensus protocols and gRPC services down at the systems layer, up through clean React/Next.js front-ends, and across into applied AI/LLM tooling. I care a lot about measurable impact — throughput, latency, and developer velocity — and about building things that are genuinely nice to use.

Data plateMeasured values
Raft txns / second
0K
imagery processed / day
0 TB
web solutions shipped
0+
graduate GPA
0
FIG. 02Experience

Where I've worked

Research, full-stack engineering, and large-scale ML infrastructure.

Exp—01

Full-Stack Software Engineering Intern

Stony Brook University — Web Dev Department · Stony Brook, NY

  • Delivered 50+ production web solutions for university departments, turning stakeholder requests into reusable React components, Node.js workflows, API integrations, and release-ready deployments.
  • Built shared component libraries, responsive layouts, form handlers, validation flows, and integration endpoints, reducing repeat implementation time by 35% across recurring launch patterns.
  • Standardized accessibility, SEO, analytics instrumentation, release checklists, and stakeholder QA notes, cutting review cycles by 30% for high-visibility communications.
[React][Node.js][TypeScript][REST APIs]
Jan 2026 — Present
Exp—02

Graduate Research Assistant — Distributed Systems

Stony Brook University — Distributed Systems Lab · Stony Brook, NY

  • Replaced Paxos replication in a geo-replicated key-value store with a C++ Raft backend, reaching 510K committed transactions/s versus 472K for Paxos in 11-worker benchmarks.
  • Separated commit from replay in a shared-log Raft design, ingesting all 11 worker lanes at 526K transactions/s while keeping follower replay within 5% of leader throughput.
  • Built parallel follower replay workers that raised throughput from 143K to 499K transactions/s and validated the design under a simulated Cloud-SSD model near 340K transactions/s.
[C++][Raft][gRPC][Distributed Systems]
Jan 2025 — May 2026
Exp—03

Software Engineering Intern — Data Infrastructure / Remote-Sensing ML

Indian Space Research Organisation (ISRO) · Ahmedabad, India

  • Engineered a parallel Python pipeline to process 1.2 TB/day of satellite imagery, cutting per-image runtime by 66% through chunked processing, reusable stages, and optimized batch execution.
  • Trained and evaluated a U-Net model for NDVI reconstruction on 1.2 TB of Sentinel-2 imagery, outperforming the Prithvi-100m baseline by 25% for vegetation and agricultural monitoring.
  • Packaged and deployed containerized batch inference with Docker and CI/CD, adding output validation and run-level monitoring to reach a 99.5% successful batch completion rate.
[Python][U-Net][Docker][CI/CD][MLOps]
Dec 2023 — May 2024
Exp—04

Internship Trainee

Matrix Comsec · Vadodara, India

  • Built an end-to-end face authentication system using PyTorch for model training and backend microservices for real-time enrollment, verification, and inference serving.
  • Developed REST API services to connect the trained authentication model with production backend systems, enabling low-latency request handling at 100+ requests/sec with sub-40 ms latency.
  • Designed a synthetic data generation and training pipeline with 10,000+ augmented face images, reducing manual labeling effort by 80% and improving model robustness.
  • Optimized the model serving path across preprocessing, inference, caching, and API layers, improving recognition accuracy by 23% while maintaining reliable service availability.
[PyTorch][Python][Computer Vision][REST APIs][Docker][Linux]
May 2023 — Aug 2023
Fig. 03 — ResearchMaster's thesis

Mako — High-Performance Raft Replication

Master's thesis: a C++ Raft backend hitting 526K committed transactions/second.

My Master's thesis at Stony Brook's Distributed Systems Lab. I redesigned replication for Mako, a geo-replicated key-value store, replacing its Paxos layer with a custom C++ Raft implementation.

Key results:

  • 526K committed transactions/s via a shared-log design that separates commit from replay.
  • Outperformed the original Paxos path (510K vs 472K txns/s) in 11-worker benchmarks.
  • Parallel follower-replay workers raised throughput from 143K to 499K txns/s while keeping followers within 5% of leader throughput.
  • Validated under a simulated Cloud-SSD model at ~340K txns/s.
0K
committed txns / second
[C++][Raft][gRPC][Shared Log][Distributed Systems]
Read the thesis
FIG. 04Projects

Things I've built

A mix of production systems, research, and side projects. Click any card for details and demos.

P—01Distributed Systems✦ Featured

Production-grade Rust node agent that runs containerized jobs over gRPC.

[Rust][gRPC][Docker][PostgreSQL]+2
P—02AI / Automation✦ Featured

AI-powered job scanner that scores roles and auto-tailors your résumé.

[Python][Flask][Claude API][GitHub Actions]+1
P—03Web App✦ Featured

A collaborative job-search tracker with a live leaderboard for a group of friends.

[Next.js][React][TypeScript][Prisma]+2
P—04Distributed Systems

A fault-tolerant background job queue deployed on AWS ECS.

[Go][Redis][PostgreSQL][Docker]+2
P—05Web App

A feature-flag control plane with rollouts, targeting, and audit logs.

[React][Node.js][PostgreSQL][Redis]+1
P—06AI / ML

A multi-agent framework for evaluating LLM prompts and models.

[Python][LangGraph][FastAPI][OpenAI API]+1
P—07AI / ML

Turns plain-English project plans into editable dependency graphs.

[React][React Flow][Claude API][Redis]+2
P—08AI / ML

A patient-centered health dashboard with statistical flare detection.

[React][Node.js][OpenAI API][Qdrant]+1
P—09Web App

A MERN food-ordering app that kills the campus canteen queue — order from your table, get pinged when it's ready.

[React][Node.js][Express.js][MongoDB]+2
P—10Web App

A Spring Boot + MVC e-commerce platform for kitchen, garden, and office products.

[Java][Spring Boot][MySQL][JSP]+2
FIG. 05Education

Where I studied

Edu—01Aug 2024 — May 2026

M.S. in Computer Science

Stony Brook University

GPA — 3.84 / 4.0

Focus on distributed systems and high-performance computing. Master's thesis: High-Performance Raft Replication for Mako.

Edu—02Oct 2020 — May 2024

B.Tech in Computer Engineering

Dharmsinh Desai University

GPA — 8.99 / 10 (Top 5%)

FIG. 06Skills

Bill of materials

From systems programming and consensus protocols to full-stack web and applied AI.

01Backend & APIs
Node.jsFastAPISpring BootDjangoExpress.jsREST APIsgRPCGraphQLWebSockets
02Cloud, DevOps & Observability
AWSGCPDockerKubernetesTerraformGitHub ActionsCI/CDLinuxNGINXPrometheusGrafana
03Databases & Storage
PostgreSQLMySQLMongoDBRedisDynamoDBElasticsearchpgvectorQdrant
04Distributed Systems
System DesignMicroservicesRaftPaxosApache KafkaEvent-Driven ArchitectureCachingLoad Balancing
05LLM & GenAI
OpenAI APIClaude APILangChainLangGraphRAGVector SearchFAISSFine-tuningPrompt Engineering
06Languages
PythonGoC++JavaTypeScriptJavaScriptRustSQLBash
07MLOps & Data
MLflowWeights & BiasesDVCPandasNumPyModel ServingExperiment TrackingData Pipelines
08Machine Learning
PyTorchTensorFlowscikit-learnKerasHugging FaceXGBoostOpenCVComputer VisionNLP
Fig. 07Contact — End of document

Let's build something together

I'm open to new-grad software engineering roles and interesting collaborations. The fastest way to reach me is email.

makadiakrish@gmail.com
Emailmakadiakrish@gmail.comGitHubgithub.com/kiiriisLinkedInlinkedin.com/in/kmakadia