OPEN TO SUMMER 2027 INTERNSHIPS

Somewhere between “what if?” and “it works.”

Kashish Phulwani - Software Developer

Software Engineer building reliable products across data, AI and distributed systems. I care about thoughtful architecture, sharp execution and technology that holds up beyond the demo.

View my workDownload resume
2.5yrs
AT JPMORGAN
6
PROD SERVICES

Introduction

About

Two and a half years shipping production backends at a global bank, now deepening the theory behind them.

I'm a software engineer with a master's in progress at CSULB and two and a half years at JPMorgan Chase, where I built RESTful services and ETL pipelines that moved millions of client records - and owned the infrastructure and release path underneath them.

Philosophy

Anyone can get a service running. The interesting question is what happens when it fails at 3am. I care about observability, reproducible infrastructure, and knowing where a system breaks before it breaks on its own.

Current focus

Retrieval systems and LLM serving - adaptive RAG routing, streaming inference, and the guardrails that keep a model from answering confidently when it shouldn't. Alongside advanced algorithms and software engineering coursework.

Selected work

Projects

Systems built end to end - retrieval pipelines, ML serving, and the architecture decisions behind them.

query
→
intent router
→
ChromaDB vectors
live web search
→
response
FeaturedAn LLM retrieval system that decides where to look before it answers.

Academic Assistant - adaptive RAG pipeline

An end-to-end LLM inference pipeline built with Python async generators, delivering responses with traceable live URL citations. Queries are routed between ChromaDB vector retrieval over academic policy data and live web search based on intent - so time-sensitive questions hit the web and policy questions hit the index.

Key features

  • End-to-end LLM inference with async generators
  • Hybrid retrieval - vector search vs. live web, routed by query intent
  • LLM query optimization layer converting conversational input to keyword strings
  • Prompt guardrails with ambiguity detection and clarifying questions
  • Traceable live URL citations on every response

Technologies

PythonChromaDBAsync I/ORAGLLM Orchestration
GitHub ↗
Machine LearningAudio in, genre out - plus a content-based recommender on top.

Music genre detection - end-to-end ML pipeline

A complete machine learning pipeline for music genre classification using KNN, covering audio feature extraction, data cleaning, model training, and evaluation. Model inference is served through a backend integrated with a responsive React interface, letting users upload audio and see real-time predictions.

Key features

  • Audio feature extraction and signal processing
  • Data cleaning, model training, and evaluation loop
  • KNN classification served through a backend API
  • React interface with real-time genre prediction
  • Content-based top-5 song recommender

Technologies

PythonKNNSignal ProcessingReactScikit-learn
GitHub ↗
CSV upload
→
quality checks
→
0-100 score
fix cost preview
→
apply one
Data QualityA readiness score that shows what each fix costs, not just what it gains.

AI-Ready Score - cost-aware data quality

Upload a tabular CSV, get a 0-100 data readiness score, and apply one fix at a time while seeing the tradeoffs. Most cleaning tools only celebrate the score going up. This one surfaces distribution shift, row loss, and new issues each repair may introduce - so the warnings are the product.

Key features

  • Quality checks for missing values, duplicates, outliers, imbalance, and more
  • 0-100 AI-Ready score with grade, severity counts, and plain-language summary
  • Cost-aware fix previews: score before/after, side effects, and verdict
  • One-fix loop - apply, re-score, re-preview instead of fix-everything
  • History, reset to original upload, and download the current CSV anytime

Technologies

FastAPIpandasNumPySDVReactVite
GitHub ↗

Capability

Skills

A backend and infrastructure stack - from APIs and data pipelines to the monitoring that proves they work.

Languages

Core programming and query languages.

PythonJavaSQLJavaScript

Backend & APIs

Services and the contracts between them.

RESTful APIsFastAPISpring BootReactGit

Cloud & Infrastructure

Reproducible environments and zero-downtime releases.

AWS (EC2, S3, Lambda)TerraformDockerJenkinsGitHub ActionsBlue-Green Deployments

Observability & Reliability

Knowing a system is healthy - and proving it survives failure.

GrafanaDatadogChaos Engineering (Gremlin)Latency Monitoring

Databases

Relational and vector data layers.

PostgreSQLMySQLChromaDB

AI/ML & Data

Retrieval systems and the data work underneath them.

RAG PipelinesLLM OrchestrationVector EmbeddingsPrompt EngineeringPandasNumPyScikit-learn

Career

Experience

Two and a half years owning backend services, data pipelines, and the infrastructure they run on.

Full-time

Software Engineer

JPMorgan Chase & Co.

Aug 2023 - Dec 2025

Built and operated enterprise backend services and data pipelines, owning everything from the API contract to the Terraform that provisioned the environment it ran in.

PythonJavaAWSTerraformDockerJenkinsGrafana

What I did

  • Designed and built RESTful APIs and backend services for enterprise data retrieval and transformation, consumed by multiple internal teams - while translating technical constraints for non-technical stakeholders.
  • Built and maintained ETL and batch-processing workflows ingesting and transforming millions of client records for downstream validation and analytics.
  • Improved observability across 6 production services with Grafana dashboards and Datadog monitors for latency, throughput, and error rates, cutting time-to-detection on incidents.
  • Built CI/CD pipelines with Jenkins and GitHub Actions using Dockerized builds, enabling zero-downtime blue-green deployments across coordinated monthly releases of all 6 services.
  • Automated AWS infrastructure provisioning with Terraform across EC2, S3, and Lambda, making every environment reproducible and auditable.
  • Developed a Python synthetic data framework (Faker) generating thousands of records across 15-20 relational tables with PII masking, enabling safe lower-environment testing without production data.
  • Helped establish the team's chaos engineering practice with Gremlin, simulating infrastructure failures to validate fault tolerance and improve service resiliency.

Academics

Education

Computer science fundamentals, from a 9.34 CGPA undergrad to a 4.00 master's in progress.

Jan 2026 - Dec 2027

M.S. Computer Science

California State University, Long Beach

GPA · 4.00 / 4.00

Relevant coursework

Advanced Analysis of AlgorithmsAdvanced Topics in Programming LanguagesAdvanced Software Engineering

Highlights

  • Perfect GPA across the first term
  • Focused on distributed systems and retrieval architecture

Aug 2019 - May 2023

B.E. Computer Engineering

University of Mumbai

CGPA · 9.34 / 10

Relevant coursework

Artificial IntelligenceMachine LearningAdvanced DBMSData Structures & AlgorithmsCloud ComputingObject Oriented ProgrammingBig Data AnalyticsSystem Security

Highlights

  • Graduated with a 9.34/10 CGPA
  • Built the systems and data foundation the JPMorgan work ran on

Recognition

Certifications & Activities

Formal credentials, and the competitions where the fundamentals got tested under a clock.

Certification

AWS Certified Developer - Associate

Amazon Web Services

Cloud-native application development, deployment, and debugging on AWS.

Hackathons

  • 24-Hour Hackathon

    JPMorgan Chase

    2023

  • Rubix Hackathon

    CSI-TSEC

    2022

  • BlindCode

    Ascent'22, TSEC

    2022

Contact

Let's build something that stays up.

Open to Summer 2027 internships and backend, infrastructure, or AI platform roles. Send a short note about your team or the role - I answer everything.

OPEN TO SUMMER 2027 INTERNSHIPS

Email

phulwanikashish@gmail.com

Location

Long Beach, CA · open to remote

GitHub

@kashish0201

LinkedIn

@kashishh-phulwani
Send me an email →

Or grab the resume first.