Python Backend Engineer with LLM Integration

ASHISH KUMAR

Python Backend Engineer Django FastAPI Flask LLM Integration

+91-831******* ****************@*****.*** linkedin.com/in/ashish-kumar-upc PROFILE

Python Backend Engineer with 4+ years building scalable microservices and enterprise APIs using Django, FastAPI, and Flask. Expert in SOLID principles, multi-tenant architecture, and database optimization. Experienced in LLM integration including OpenAI, Ollama (Llama), RAG systems with vector databases, and LangChain workflows. Proficient with Claude, Codex, and AWS cloud services. Proven track record of 25-35% performance improvements through architectural optimization. TECHNICAL SKILLS

Languages & Frameworks: Python, Django, FastAPI, Flask, Django REST Framework, GraphQL, SQLAlchemy

Backend & Architecture: REST APIs, GraphQL, JWT Authentication, SOLID Principles, Repository Pattern, Microservices, Event-Driven Architecture, Dependency Injection Database & Performance: PostgreSQL, Query Optimization, Strategic Indexing, N+1 Prevention, Connection Pooling, EXPLAIN ANALYZE

Cloud & DevOps: AWS (ECS, EKS, CodeBuild, Lambda, API Gateway, Cognito), Docker, Kubernetes, Terraform, GitHub Actions, ArgoCD, CI/CD

Tools: Claude, Codex, Kiro, Git, Jira, MQTT, Alembic, Uvicorn LLM & Generative AI: OpenAI API, Ollama, LangChain, ChromaDB, HuggingFace Embeddings, Llama, RAG Systems, Function Calling, Tool Use, Prompt Engineering, BeautifulSoup, Gradio, Pydantic (Structured Outputs)

EXPERIENCE

Software Developer — Aloha Technology Pvt Ltd., Pune, India Feb 2022 - Present (4+ Years)

• Architected multi-tenant platform with SOLID principles, layered architecture, and dependency injection; deployed on AWS EKS with rate limiting via Ingress annotations.

• Delivered 20+ REST APIs (Flask-Smorest, SQLAlchemy) with JWT + AWS Cognito RBAC, improving response time from 800ms to 600ms (25% improvement).

• Optimized PostgreSQL queries using strategic indexing and N+1 prevention, reducing API latency by 35% under peak load (10,000+ req/min).

• Implemented event-driven architecture with MQTT for asynchronous task processing and CI/CD automation with AWS CodeBuild, reducing deployment time by 30%.

• Debugged critical N+1 query issues (150+ to 3 calls) and deadlock conditions, increasing job completion rate from 85% to 98%.

KEY ACHIEVEMENTS

• Implemented rate limiting via Kubernetes Ingress annotations supporting 10,000+ requests/minute with <1% failure rate, preventing DDoS attacks and optimizing infrastructure costs.

• Automated AWS infrastructure provisioning with Terraform, reducing manual deployment overhead by 40% and enabling Infrastructure-as-Code best practices across all projects. PROJECTS

Weave — Multi-Tenant Workflow Platform

Flask SQLAlchemy PostgreSQL Docker AWS EKS GitHub Actions MQTT

• Built multi-tenant SaaS with tenant-level data isolation, RBAC, and event-driven workflow engine; achieved 25% latency reduction on critical endpoints.

• Implemented rate limiting via Kubernetes Ingress annotations (10,000+ req/min), JWT + AWS Cognito authentication, and asynchronous email processing with template rendering.

• Applied repository pattern, service-oriented architecture with dependency injection, and Infrastructure-as-Code using Terraform.

Metadoc AI — Document Processing Platform

Django REST GraphQL PostgreSQL MongoDB AWS Lambda ECS

• Architected hybrid REST + GraphQL API system with microservices pattern; improved response times by 30% through efficient query optimization.

• Configured AWS Lambda serverless processing with S3 integration for scalable document storage and MongoDB for flexible schema management.

• Designed scalable API Gateway CRUD operations with IAM role policies and Lambda Authorizer for enterprise-grade authentication, ensuring secure multi-tenant document access. Copper Connector — Real-Time Data Sync

FastAPI SQLAlchemy PostgreSQL Docker MQTT Uvicorn

• Built high-performance REST API using FastAPI with dependency injection pattern; implemented MQTT publish-subscribe for asynchronous message processing.

• Secured API with JWT authentication; containerized with Docker Compose and managed schema versioning via Alembic migrations with scheduled cronjobs for background tasks. Page Buddy — AI-Powered Web Chat Assistant

Python Ollama (Llama 3.2) BeautifulSoup Gradio OpenAI-compatible Client

• Built a web chat app where users provide a URL; scraped and parsed page content using BeautifulSoup, then injected it as context into a local Llama 3.2 model served via Ollama.

• Integrated Ollama via OpenAI-compatible client (custom base URL + API key) for local LLM inference; built interactive chat UI with Gradio and implemented function calling to fetch live product pricing.

AI Job Matcher — Resume-to-Job RAG System

Python LangChain ChromaDB HuggingFace Embeddings Llama OpenAI API

• Developed a RAG-based job matching system where users upload a resume and receive semantically matched job listings using ChromaDB vector store and HuggingFace embeddings.

• Leveraged Llama and OpenAI models via LangChain for flexible LLM switching; used Pydantic for structured output parsing and prompt engineering for accurate skill extraction. EDUCATION & CERTIFICATIONS

Master of Computer Application (MCA) — Dr. APJ Abdul Kalam Technical University 8.3/10 GPA 2022

Bachelor of Computer Application (BCA) — Mahatma Gandhi Kashi Vidyapeeth University 68% 2020

Certifications

• Kubernetes with Amazon EKS, Fargate & DevOps AWS Lambda & Serverless Kubernetes From Scratch Python Django for AWS AI Engineer Core Track: LLM Engineering, RAG, QLoRA & Agents

Languages: English (Fluent) Hindi (Fluent)

Contact this candidate