kernel space / systems console
This is not a portfolio wall. It is a technical dossier for a builder working across data infrastructure, backend engineering, and applied AI.
M.S. Applied Data Science @ San Jose State University
Open to backend, data, and platform roles
Bias toward systems that survive real-world use
Uptime
99.987%
Latency
14ms
Ingest Rate
1.24 GB/s
Live build surface
Currently building
Agentic workflows, graph-based reasoning, and systems that feel product-grade instead of experimental.
Builder habits
Midnight coding, architecture diagrams before implementation, and making complex systems readable to both engineers and non-engineers.
What I optimize for
Clarity under pressure, strong operational defaults, and technical work that still looks good after real-world traffic, failure, and change.
Systems OS
This is the part that should feel different. Switch modes and the system surface changes with the kind of work being done.
Mode 01
This mode is about boundaries, sequencing, and turning vague product pressure into a system shape that can actually ship.
Active case file
Enterprise RAG
Enterprise retrieval, auth, caching, and vector infrastructure delivered as one product-grade system.
Open case fileApplied AI system
Enterprise-grade vector search, auth, and caching infrastructure.
Identity chamber
I do not want to be remembered as someone who can list tools. I want to be remembered as someone who can structure complexity.
The work I care about sits at the intersection of infrastructure, intelligence, and product clarity.
The best systems are not only fast or clever. They are legible, observable, and trusted by the people using them.
Throughput gain
0%
Refactored backend services from O(n²) to O(n log n) throughput using B+ trees and hashing on Linux services.
p95 latency drop
0%
Achieved via intelligent database indexing, query optimization, and read-replica replication schemes.
Records processed daily
0M+
Successfully handled in high-availability web and ETL platform configurations using React, FastAPI, Postgres, and Redis.
Release time cut
0%
Cut deployment overhead with automated CI/CD pipelines, blue-green releases, and zero-downtime rollouts.
Medium signals
This section shows the public thought process behind the systems work, so visitors see more than repos and buzzwords.
Systems Engineering
A deep dive into rewriting database indexing structures and resolving query latency drag on high-throughput backend services.
Platform Operations
My operational checklist for scheduling distributed jobs, Celery executors, and configuring EKS node pools to optimize wall-clock efficiency.
Applied AI Design
Connecting autonomous supervisor routing nodes to zero-hallucination vector retrievals and generative transaction payments.
Product Infrastructure
How we scaled the award-winning SJ HOPES platform to handle concurrent traffic spikes during emergency shelter allocations.
Applied AI Design
Architecting autonomous travel recovery assistants with strict term filters, context checking, and fallback search models.
DevOps & SRE
Building automated logs analysis systems and setting SLO threshold indicators to decrease production incidents.
Experience log
Not a list of titles. A trace of environments where scale, reliability, and technical ownership were the daily operating conditions.
Industry
MeteoControl
Industry
Dupat Infotronicx Pvt. Ltd.
Education signal
Degrees are context, not credentials. These programs shaped how I think about data, systems, and applied intelligence.
Master of Science
San Jose State University
Relevant coursework
SJ Hacks 2026 Winner — civic-tech platform with 1000+ user target
MongoDB Agentic Hackathon Finalist — multi-agent travel OS
LangGraph contributor — Agentic Commerce architecture pattern
Bachelor of Technology
Silver Oak University
Relevant coursework
Led software engineering group projects
Top academic performance in CS track
Stack atlas
Most portfolios dump tools into badges. This section maps how the technical language of the work bends toward architecture, scale, intelligence, and shipping.
lane 01
foundation
Python
Java
Go
C++
TypeScript
SQLlane 02
intelligence
lane 03
interfaces
FastAPI
Spring Boot
Node.js
React
Next.js📡gRPC🔌WebSocket
NGINXlane 04
data flow
PostgreSQL⚡AWS Aurora💾DynamoDB🍃MongoDB Atlas
Redis🔀Kafka
Spark🌬️Airflowlane 05
shipping
AWS
Docker
Kubernetes🚀GitHub Actions
Jenkins
Prometheus
Grafana
Jiralane 06
ml systems & scale
PySpark
Hadoop📈MLflow🔍Voyage AI🍃MongoDB Vector Search
Scikit-learn
Pandas
NumPySkills constellation
Hover to explore connections. Node size reflects depth. Edges show how tools relate in real projects.
Operating model
Instead of showing a generic toolbox list, this section shows how I approach technical work from inputs to deployment.
Ingest
I start by understanding sources, constraints, and failure patterns. Clean systems begin with honest inputs, not optimistic assumptions.
Tools in this mode
Proof of use
Used in ETL design, data platforms, graph fraud workflows, and multi-source forecasting pipelines.
Case files
Each card now opens with a stronger claim. Inside, the project pages carry the architecture, tooling, impact, and delivery story.
Applied AI system
A high-performance retrieval engine for enterprise knowledge workflows.
Enterprise-grade vector search, auth, and caching infrastructure.
Stack + delivery
FastAPI, React, pgvector, Redis, Docker, and JWT auth.
Graph ML / fraud intelligence
Graph-based intelligence pipeline exposing relational fraud rings.
Expose laundering rings and coordinated transaction abuse.
Stack + delivery
Neo4j, Graph Data Science, Python, and analyst dashboard.
Platform engineering
ETL infrastructure built with quality rules and autoscaling discipline.
5 TB+/day ETL pipeline built for scale and operational quality.
Stack + delivery
Airflow, Spark, Kubernetes, Prometheus, and Grafana.
Agentic product system
Multi-agent travel OS with vector search and Coinbase CDP.
Context-aware traveler recovery system with persistent memory.
Stack + delivery
LangGraph, Atlas Vector Search, FastAPI, and Next.js.
Data governance system
Automated metadata catalog for discovery and compliance.
Governance infrastructure turning tribal knowledge into system memory.
Stack + delivery
Python, custom catalog flows, and metadata automation.
Big Data ML / Trust Systems
Big Data review filter and trusted recommender engine.
6-layer multi-modal defense against bot farm manipulation.
Stack + delivery
PySpark, Spark MLlib, K-Means, ALS, and Streamlit.
Machine Learning / EDA
Predictive modeling pipeline for pre-release content success.
Predict pre-release content success using synopses and metadata.
Stack + delivery
Python, scikit-learn, Random Forest, and Power BI.
Deep Learning / Audio DSP
Spatial-temporal deep learning speech classifier.
Detect human emotional states from spoken audio in real-time.
Stack + delivery
PyTorch, Librosa DSP, 2D CNN, Bi-LSTM, and Attention.
Build graph pipeline
Hover over nodes to explore upstream sources and downstream effects. Observe data flows moving through the architecture.
Orchestrates batch ETL processing & workflows.
Tunes data partitioning & processing of 5TB+/day.
Ingests real-time telemetry & system event streams.
High-performance relational DB storage.
Schedules pods, autoscaling, and service meshes.
Stores raw parquet & json streams for batch processing.
Managed Kubernetes for scalable spark & model services.
Delivers async endpoints with sub-100ms P95 latency.
Powering enterprise microservices & Java logic.
High-speed internal microservice RPC communication.
Handles real-time bi-directional server sync.
Load balancing, reverse proxies & entry routing.
Designed endpoints with rate-limiting & auto-validation.
Multi-agent autonomous systems using state graphs.
Vector search retrieval pipelines to reduce hallucination.
Deep learning models for forecasting & classification.
High-dimensionality index matching on MongoDB/Atlas.
Packaging & deploying weights to online APIs.
Feature extraction, plot synopsis tokenization.
Random Forest popularity classifiers & regressions.
Rapid prototyping under high-pressure windows.
Interfaces built for generative AI and agent states.
Making deep technical concepts readable & compelling.
Optimizing startup, page transitions & DB calls.
Building responsive tools with immediate feedback.
Interactive dashboards for data quality & execution.
Business insights and popularity analytics reports.
GitHub pulse
Contribution heatmap from the last 6 months. Consistency is the strongest signal of a working builder.
Contributions
847
PRs Merged
142
Total Stars
84
Global Rank
Top 4.2%
Current Streak
12d
Longest Streak
34d
Active Days
32
Core Libs
LangGraph
Local Telemetry Sandbox
Click on the circle nodes (Center, Left, or Right) to toggle their routing direction!
diagnostics engine
A simple packets pipeline test simulation. Change router states dynamically to match database, cache, and service requests.
Contact vectors
So this section does not pretend every conversation starts the same way.
Recruiter path
For hiring conversations around backend, data, platform, or applied AI roles.
Email Arya
Builder path
For collaborators, founders, and teams who want someone strong in systems and execution.
Open LinkedIn
Proof path
For people who want code, repos, experiments, and implementation signal first.
Open GitHub