Prachi Sharma

Software Engineer

Ghaziabad 1+ yrs exp 82 · Excellent

About

Software Engineer with 1+ year of experience building production-grade backend systems and generative AI applications using Python, Django, and GCP. Designed and deployed a two-agent AI orchestration layer (Claude tool-use API) and an automated scheduling engine for a large-scale manufacturing ERP. Experienced in scalable distributed systems, concurrency-safe transactional APIs, and taking AI-powered solutions from concept to deployment. Strong DSA foundation in C++ (top 8% GeeksforGeeks, 5,211 participants).

Skills & Expertise (28)

Generative AI Intermediate

7.5/10

Years Exp

JavaScript Intermediate

6.3/10

Years Exp

Git Intermediate

6.3/10

Years Exp

Google Cloud Platform Intermediate

6.3/10

Years Exp

REST APIs Intermediate

6.3/10

Years Exp

Python Intermediate

6.3/10

Years Exp

Django Intermediate

6.3/10

Years Exp

Node.js Intermediate

5.5/10

Years Exp

Docker Intermediate

5.5/10

Years Exp

GraphQL Beginner

4.5/10

Years Exp

GitHub Express.js C++ SQLite Postgresql MongoDb MySql bootstrap Tailwind CSS Typescript React.js Claude API Prompt Engineering RAG Pipelines LLM Orchestration Async programming Redis Celery

Work Experience

SDE-1

Likwid Techserve Pvt. Ltd.

Jun 2025 - Present

Architected and shipped a two-agent LLM orchestration layer (Production Agent + Procurement Agent) on a Django/GAE ERP using Claude's tool-use API and structured prompt engineering, replacing hardcoded scheduling logic with intelligent reasoning over inventory, machine availability and order priority; with zero-regression fallback across 100% of existing accounts. Engineered a BOM-driven production scheduling engine for 1000+ machines, dynamically allocating jobs across machinery pools based on real-time machine availability, material inventory and turnaround-time constraints, cutting manual planning effort by ~60%. Delivered Delivery Order APIs using Django atomic transactions and select_for_update for FIFO batch deduction, releasing stock commitments and posting dispatch in a single atomic call; replacing 4 manual DB steps per order and eliminating a critical race condition in post-production order state management. Built and deployed Cogent backend infrastructure (AI-powered website builder) on GCP, including a custom domain provisioning pipeline, zero-touch SSL certificate provisioning via GCP's native SSL management, and a support ticketing system with email notifications; onboarding 50+ users in production. Implemented a machinery maintenance automation system for 500+ machines using Django, Celery, Redis and cron-driven management commands with time-based and output-based triggers, reducing scheduling overhead by ~40% and eliminating manual Excel/WhatsApp workflows. Designed a priority-scoring engine weighting delivery urgency (50%), customer LTV percentile (30%) and order value percentile (20%) to auto-sequence high-value orders into earlier machine slots, removing manual triage across 500+ orders/month. Identified and resolved a critical scheduling gap where production plans silently stalled on material shortages; implemented auto-rescheduling on stock receipt via Django post_save signals and introduced EOQ-based vendor ranking into procurement recommendations. Shipped an RFQ module (Django REST + React + TypeScript) with vendor-inventory mapping, single/multi-vendor RFQ creation and HTML email quotation tracking; replacing a bulk-Excel-only workflow with one-click outreach to 10+ vendors per item and reducing RFQ cycle time by ~65%.