Pratik
Sontakke
AI Engineer specializing in production-ready RAG systems, autonomous Agents, and PEFT fine-tuning—delivering reliable, explainable AI powered by Python, LangChain/LangGraph, and AWS.
I build secure multi-tenant backends, craft retrieval pipelines that surface the right context, and ship resilient, low-latency APIs with strong observability. Experienced across FastAPI, Spring Boot, pgvector/Pinecone, and CI/CD pipelines that keep systems scalable and dependable.
Technical Skills
AI & Machine Learning
LLMs, RAG, AI Agents, Fine-Tuning (PEFT), LangChain, LangGraph
Backend & APIs
FastAPI, Spring Boot, Microservices, REST APIs
Languages & Databases
Python, Java, JavaScript, SQL • pgvector, PostgreSQL, MySQL, Redis • Pinecone, Chroma
Cloud & DevOps
AWS, Docker, Kubernetes, Terraform, GitHub Actions, Vercel
Projects
Multi-Tenant RAG-as-a-Service Platform
Production-grade conversational AI platform with secure schema-per-tenant isolation and RBAC, projected to save clients $20k and 1–2 months of build time.
- FastAPI backend on AWS with real-time, context-aware responses
- PostgreSQL with schema-per-tenant and fine-grained RBAC
Intelligent RAG Assistant (Google Workspace)
Multi‑modal assistant that centralizes knowledge workflows and automates email—boosting productivity by ~20% and cutting research time by up to 75%.
- Ingests Google Drive files and drafts Gmail with OpenAI models
- n8n-orchestrated pipeline across OpenAI, PostgreSQL/pgvector, Workspace
Professional Experience
Senior Software Engineer AI - 5C Network
- Developing an end-to-end workflow tool that automates model training from annotated datasets, enabling rapid, repeatable experiments and analysis for ML team leads.
- Engineered an AI agent to automatically verify and validate over 5,000+ radiology reports daily, decreasing manual review time by 30 % and enhancing report reliability.
- Developed and deployed a RAG model for the company’s landing page to automate user query responses, reducing the workload for the technical support team.
Cloud & AI Backend Engineer - Freelance
- Engineered a Text-to-SQL agent using LangChain and integrated it into a Spring Boot service, empowering non-technical teams with self-service analytics and saving 5-10 hours weekly.
- Architected core cloud infrastructure on AWS using Terraform and established a full CI/CD pipeline, slashing service deployment times from hours to under 10 minutes.
- Developed and containerized key Java/Spring Boot microservices to serve AI models, ensuring a resilient and responsive backend system capable of handling production traffic.
Software Engineer - Guenstiger
- Optimized high-traffic Java/Spring Boot services by refactoring the data access layer and re-architecting core components, cutting API response times by an average of 200ms.
- Led backend enhancements that improved system throughput by 15 %, supporting a growing user base without additional infrastructure costs.
Junior Technical Consultant - Edifition
- Translated client business needs into detailed technical specifications for 5+ projects, leading to a 15 % reduction in requirement-related change requests post-development.
Education
AI Engineering
Bachelor of Science (B.Sc), Computer Science
Let's Connect
Open to roles in Fine-tuning (PEFT), ML, AI Engineering, and collaborations on RAG/Agents projects.