Active Systems Engineer • Open to Roles

Building Production-Grade AI Systems & High-Scale Applications

I design and implement custom LLM pipelines, multi-agent orchestrations, and robust database layers. Focused on latency optimization, token-cost reduction, and rigorous systems architecture.

Get in Touch GitHub Profile

98%

RAG Accuracy Target

-38%

LLM Token Savings

<120ms

DB Retrieval Speed

rag_system_monitor.py

LIVE

$ python -m evaluator.pipelines --target production --cost-opt

[11:24:03] INFO: Inbound query: "Compare leverage metrics in financial statements"

[11:24:03] EMBED: Tokenizing string & calling text-embedding-3-small -> OK [18ms]

[11:24:03] RETRIEVE: Querying Milvus Vector DB (top_k=7, similarity_threshold=0.84)

[11:24:04] RETRIEVE: Found 7 relevant document vectors across 3 reports -> OK [35ms]

[11:24:04] COMPRESS: Invoking custom LlamaIndex context compression parser

[11:24:04] COMPRESS: Compressed prompt context window size by 54.8% (12,400 tokens -> 5,600 tokens)

[11:24:04] ROUTE: Latency-optimized LLM router directed workflow to Anthropic Claude 3.5 Sonnet

[11:24:05] LLM: Pending API resolution from anthropic.sonnet-v2-eval...

[11:24:05] LLM: Received 480 token response -> OK [620ms]

[11:24:06] GUARD: Running validation loops: Relevancy Score (0.98), Factuality (0.99)

[11:24:06] SUCCESS: Payload dispatched to UI in 673ms. LLM Token Cost: $0.00375

CPU load: 1.4%

API Status: 200 OK

Latency Optimized: Enabled

ENGINEERING ARCHIVE

Featured Production Systems

Deployed, high-performance systems demonstrating full-stack engineering competency, system scalability, and practical AI implementations.

Manufacturing · B2B SaaS Phase 1 · Raw Material Inventory

Prodexa

AI-Ready Smart Manufacturing Operations Platform

A full-stack B2B SaaS platform that centralizes raw material inventory, supplier operations, and manufacturing workflows for small and medium manufacturers. Phase 1 delivers a structured digital foundation replacing fragmented manual systems.

Current Capabilities — Phase 1

Auth & RBAC Material Management Supplier Management Stock In / Out Stock Adjustments Dashboard Audit Logs

Engineering Highlights

• Modular Next.js + NestJS monorepo architecture with TypeScript across the full stack
• PostgreSQL-backed inventory workflows with Prisma ORM and schema migrations
• JWT authentication with role-based access control (RBAC) for enterprise multi-user support
• REST API layer with Swagger / OpenAPI documentation for structured integration
• Structured audit log system for full operational traceability and inventory accountability

AI Roadmap — Planned Intelligence Layer

Demand Forecasting AI Inventory Planning AI Operational Insights Production Optimization LLM Integration

These capabilities are part of the future product roadmap and are not yet implemented.

Next.js NestJS TypeScript PostgreSQL Tailwind CSS

Case Study GitHub

FinTech · Banking AI-Assisted Qualification

LeadNexus

A full-stack FinTech platform for personal loan lead qualification, behavioral tracking, and sales workflow orchestration — helping banking teams identify and manage higher-potential loan prospects through a centralized, data-driven system.

Engineering Highlights

• Full-stack Next.js + NestJS architecture with TypeScript across the entire frontend and backend codebase.
• Zustand-powered frontend state management with Recharts for lead analytics and sales pipeline visualization.
• Groq API integrated for AI-assisted lead qualification and scoring support within the sales workflow.
• PostgreSQL-backed relational data layer with Prisma ORM and structured migrations for lead and workflow persistence.

Next.js NestJS TypeScript PostgreSQL Groq

Case Study GitHub

LEAD PIPELINE · OVERVIEW ACTIVE

High 24 Leads

Medium 61 Leads

Low 38 Leads

Qualified

19 leads

In Review

34 leads

Contacted

28 leads

AI-assisted scoring via Groq API · Qualification support

Full-Stack / Automation Metrics: 0% Stockouts

MediStock AI

Built an event-driven stock monitoring and expiration forecasting engine for pharmacies. Automates manual drug expiration sheets by dynamically analyzing levels and executing auto-replenishment notifications.

Engineering Highlights

• Developed a high-throughput event processing backend with FastAPI background workers, resulting in real-time alert dispatches.
• Integrated Groq-based inference pipelines to calculate complex drug expiration rates and suggest optimal replenishment levels.
• Enforced rigorous ACID transaction isolation constraints inside SQLite database layers to prevent race conditions during concurrent batch orders.

Python FastAPI Groq LLM SQLite

Source Code

REAL-TIME INVENTORY MONITORS ONLINE

API Response Time <85ms

Expiration Precision 96.8% Match

Simultaneous Requests 1,200 req/sec

Full-Stack Web App Vercel Deployed

BookTracker Platform

A highly reactive full-stack web dashboard that helps readers log progress, manage book queues, and parse custom data from the Google Books API.

Engineering Highlights

• Built a highly responsive Vue 3 client-side state architecture, maintaining client-sync and instant rendering cycles.
• Designed a fully indexed relational SQLite database schema, avoiding heavy subqueries to compute speed and completion rates.
• Implemented an asynchronous backend cache on Google Books API requests, dropping average request time overhead by 6x.

Vue 3 FastAPI SQLite Google Books API

Source Code Live Demo

API RESPONSE METRICS CACHED

Google API Raw Fetch 720ms

Local Cache Hit Fetch 115ms (6.2x Faster)

Green represents latency reduction profile.

INFRASTRUCTURE CAPABILITY MAP

Structured Technical Capabilities

System and stack layers structured by engineering discipline, highlighting high-level competencies built over years of development.

AI & RAG Systems

Design and implementation of vector ingestion layers, semantic retrieval routers, dense representation embeddings, and guardrail validations.

LangChain LlamaIndex Vector Search Multi-Agents Prompt Compression

Backend & API Systems

Constructing event-driven RESTful endpoints, asynchronous background processing workers, server architectures, and API middleware schemas.

FastAPI Node.js Spring Boot REST APIs Microservices

Databases & Search

Data modeling across relational, analytical time-series, and multidimensional vector namespaces. Query routing optimization and index patterns.

PostgreSQL ClickHouse Milvus DB Pinecone SQLite

Cloud & Infrastructure

Configuring multi-container environments, automating pipeline deployment workflows, cloud operations, and static CDN assets.

Docker AWS ECS GCP CI/CD Pipelines Vercel / Git

Frontend & Interface

Structuring highly responsive client-side UI layouts, component architectures, atomic CSS, and performant state synchronization.

Next.js Vue 3 React.js Tailwind CSS State Sync

Programming Languages

Proficient across object-oriented and analytical languages, utilizing functional structures, type definitions, and algorithmic procedures.

Python SQL JavaScript (ES6+) Java Bash

WORK HISTORY

Professional Engineering Experience

A chronologically detailed path of systems architecture, optimization metrics, and database scalability achievements.

Software Engineer

ACTIVE ROLE

2026 — Present

Talent Smart • Series-A AI Ingestion Architectures

Lead developer architecting multi-agent processing systems and dense document context summarization engines for high-performance retrieval requirements.

▶ Architected custom LlamaIndex pipelines supporting hybrid vector/keyword database routing, target indexing 98% factuality retrieval rates.
▶ Built semantic compression handlers inside agent workflow structures, lowering API token expenditure rates by 38% while improving latency by 45%.
▶ Integrated distributed indexing structures across Pinecone and Milvus vector spaces, maintaining high-throughput lookups for multi-user client apps.

Software Engineer Intern

2025 — 2026

AlpineCode Technology • Time-Series Analytics & ClickHouse Optimization

Engineered and optimized backend relational database schemas and automated processing pipelines storing high-frequency data streams.

▶ Optimized analytical storage architectures on ClickHouse, structuring custom database partition keys to capture millions of metric event flows.
▶ Engineered optimized analytical database SQL scripts, resulting in a 4x runtime speedup on multi-million row aggregation datasets.
▶ Constructed reliable data retrieval streams integrating analytical outputs directly into Express/Node.js API endpoints.

Advanced CS Foundations & Systems Theory

2023 — 2025

Ace Academy • Rigorous Graduate-Level CompSci Preparation (GATE CSE)

Dedicated 24 months to intensive theoretical computer science coursework, mastering low-level system designs, mathematical computation, and algorithmic efficiency profiles.

▶ **Data Structures & Complexity Algorithms**: Engineered optimal solutions solving graphs, memory structures, dynamic tree optimizations, and complexity limits.
▶ **OS Internals & Concurrency Control**: Evaluated process dispatch algorithms, deadlock resolutions, virtual memory architectures, and atomic socket operations.
▶ **Relational Database Theory**: Mastered normalization levels (1NF-BCNF), transactional isolation guarantees, and file storage tree architectures (B+ Trees).

Programmer Analyst Trainee

2022 — 2023

Cognizant Technologies Solutions • Client-Side Systems & Component Layouts

Collaborated in Agile teams building responsive web application components and client UI systems utilizing utility-first CSS frameworks.

▶ Created atomic, modular UI React components integrated with Tailwind CSS systems, improving client layout loading times.
▶ Built stable client state synchronization scripts, eliminating rendering lag issues and increasing web-vital scores by 25%.
▶ Participated in standardized codebase sprint releases, sprint specs, code reviews, and structural audits.

GenC Intern

2022

Cognizant • Full-Stack Microservices Ingestion

Contributed to internal portal development, integrating secure React interfaces with Spring Boot microservice handlers.

▶ Co-developed and tested 5+ Spring Boot REST endpoints, securing request routing channels.
▶ Built microservice Dockerfiles and verified local runtime container configurations.
▶ Configured basic build triggers on AWS ECS staging systems to support team deployment needs.