Rudresh Naik — Backend Software Engineer

About

Backend engineer with 5+ years designing, building, and operating production systems for a US LegalTech company — the lead-acquisition, scoring, routing, and partner-integration platform its business runs on. Deep in Laravel/PHP, REST APIs, Elasticsearch, MySQL/BigQuery, and Google Cloud; working proficiency in Python and JavaScript; ships AI-integrated products end-to-end (speech pipelines, local LLMs, vision models).

Skills

PHP Python JavaScript SQL Laravel FastAPI REST APIs Queues & Background Jobs Redis Linux MySQL Elasticsearch BigQuery Whisper (STT) Ollama F5-TTS Google Cloud Platform Docker Git Stripe Twilio DocuSign SendGrid Google Ads API React Next.js Claude Code

Experience

Backend Software Engineer

March 2021 – June 2026

accident.com — US LegalTech, remote

Built and maintained the production Laravel platform powering lead acquisition, qualification, routing, and distribution for US law firms — the company's core revenue system.
Designed and shipped REST APIs for lead ingestion, partner integrations, and business workflows connecting marketing sources to law-firm buyers.
Built Elasticsearch-powered search and lead-scoring systems across large production datasets, improving match quality and routing speed.
Developed analytics pipelines on BigQuery that powered reporting and campaign decisions for US stakeholders.
Integrated Stripe, Twilio, DocuSign, SendGrid, and the Google Ads API into revenue-critical production workflows.
Operated services on Google Cloud Platform with Docker (Cloud Scheduler, Secret Manager); profiled and optimized MySQL queries and backend hot paths for reliability under production load.
Collaborated directly with US stakeholders in English for five years — requirements, delivery, and production support.

PHP Laravel Elasticsearch Google Cloud BigQuery Google Ads API Twilio Stripe DocuSign SendGrid REST APIs MySQL JavaScript

Projects

Voice Buddy — offline AI Mandarin tutor

Python · FastAPI · github.com/ruddesh/voice-buddy

Fully offline speech-practice platform: faster-whisper speech-to-text, scenario-based conversation with Qwen 2.5 (14B) served via Ollama, F5-TTS neural voice cloning, and per-syllable Mandarin tone scoring built on Praat pitch-contour extraction; includes SM-2 spaced-repetition drills.

Python FastAPI Whisper Ollama F5-TTS

Textbook OCR — vision-LLM vocabulary pipeline

Python · Ollama · github.com/ruddesh/textbook-ocr

Turns photos of Chinese textbook pages into validated flashcards: a local vision-language model (Qwen2.5-VL) extracts vocabulary; a normalization pipeline converts Simplified to Traditional, corrects pinyin against ground truth, and de-duplicates entries; a second LLM pass generates mnemonics and example sentences.

Python Ollama Qwen2.5-VL OCR

Education

Full-time Mandarin Language Program — Taipei, Taiwan · 2026 – present

Undergraduate Engineering Coursework (2 years) — India

Languages

English (fluent — 5 years of daily professional use) · Hindi (native) · Konkani (native) · Mandarin Chinese (elementary, studying full-time)