Headline

Software Engineer at Cargofirst | Backend & AI Engineer | RAG, LLM Eval, Distributed Systems | Ex-HCLTech

About

I am a software engineer driven by elite technical execution, specializing in production AI systems and scalable backend architecture. Holding a Top 1% global algorithmic rank as a CodeChef 5-Star and Codeforces Expert, I thrive on solving complex, high-pressure engineering challenges and shipping functional products fast. ​Currently, as a Founding Engineer at Takkada, I am architecting event-driven Python backends and custom multi-agent orchestration systems (LangGraph, CrewAI) for AI-powered financial automation. I also recently built Ibhaveda.com from scratch as the sole Founding Engineer, where my core system design and technical execution led directly to the platform's selection into the prestigious, globally selective Narwhal Ventures cohort in Amsterdam. ​Core Technical Deliverables: ​Autonomous Systems: Building agentic workflows with shared memory, tool-calling, and eval harnesses to handle complex multi-tenant environments. ​Performance Optimization: Fine-tuned LLMs on Azure ML (TinyLlama + LoRA/PEFT) to reduce inference latency by 35% and optimized robust RAG pipelines (pgvector) to achieve 85% semantic accuracy and a 90% reduction in manual review. ​Scalable Architecture: Designed high-throughput Python/FastAPI and Node.js backends equipped with async queues and state machines to manage concurrent workloads. I recently architected a real-time SaaS platform handling infinite-depth idea branching with millisecond-latency synchronization. ​High-Velocity Execution: Won 1st Place at the WCE Hackathon and reached the National Finals at the Smart India Hackathon (SIH), proving an elite ability to build and deploy functional MVPs on ultra-tight deadlines. ​Previous Experience: ​Software Engineer (AI & Backend) @ HCLTech: Built conversational agents and automated CI/CD pipelines, cutting deployment times by 40%. ​Full Stack Engineer @ Adaapt.ai: Optimized PostgreSQL and production RAG pipelines to deliver a 3x increase in system throughput. ​Top 15% CS Graduate @ VIT | Open-source contributor to Tavily AI | Former Web Dev Lead for a 5L+ user portal. ​I am actively seeking GenAI, AI Backend, or Agentic AI roles in India or remotely. If your team values execution speed, robust infrastructure, and building reliable AI products from scratch.

Location

Bangalore Urban

AI stack

claude code

langchain

langgraph

livekit

openai

huggingface

Work experience

AI Engineer at Emergent, Jul 2026 - Present · 2 mo, Bengaluru

Co-Founder & Chief Technology Officer (CTO) at Piing AI, Jul 2026 - Present · 2 mos, Bengaluru, • Led the technical vision, AI architecture, infrastructure, and full-stack engineering for piing, an India-first AI matchmaking platform built natively on WhatsApp. • Architected and deployed a scalable serverless backend using InsForge (PostgreSQL) with Row-Level Security (RLS), seamlessly integrated with the Meta WhatsApp Cloud API to support high-throughput conversational AI. • Engineered a low-latency multi-LLM intent extraction pipeline combining Groq (Llama 4 Scout) and Gemini 3.1 Flash-Lite with intelligent API key rotation, delivering structured intent extraction from natural language conversations in sub-second latency. • Built a real-time serverless graphics engine using Deno, TypeScript, Satori, and resvg-wasm to dynamically generate personalized match cards and stream PNG images directly into WhatsApp conversations. • Designed a zero-trust privacy architecture with cryptographic state management and double-opt-in consent workflows, ensuring user identities and phone numbers remain protected until mutual matches are confirmed. • Developed a high-performance Next.js 15 frontend featuring advanced UI interactions, including Framer Motion animations, blueprint grid layouts, cursor spotlight effects, fluid SVG typography, scroll-linked animations, and viral referral systems. • Implemented resilient webhook processing with custom message deduplication, automated retry handling, secure WhatsApp System User token management, and fault-tolerant infrastructure to ensure reliable, uninterrupted platform operations.

Back End Developer at Takkada, Apr 2026 - Present · 5 mos, Remote, • Architected event-driven, queue-based Python backend services for AI-powered financial automation workflows. • Designed a database-driven state machine with atomic RPC transitions for automated multi-step business onboarding. • Built stateless webhook pipelines with comprehensive observability systems ensuring production reliability.

Python developer at HCLTech, Apr 2025 - Apr 2026 · 1 yr 1 mo, Lucknow, • Fine-tuned Large Language Models (TinyLlama) on Azure ML, achieving 85% task accuracy. • Designed and deployed production RAG pipelines and multi-turn conversational agents using LangChain. • Built and maintained CI/CD pipelines with Docker & Jenkins, reducing deployment cycle time by 40%. • Integrated Azure Cognitive Services to track model quality and minimize hallucinations.

Python Developer at Adaapt AI, May 2024 - Apr 2025 · 1 yr, Bengaluru, • Developed end-to-end Generative AI applications by integrating OpenAI GPT APIs with FastAPI backend services, managing over 500 daily interactions. • Built production-grade RAG pipelines using pgvector, achieving 85% semantic accuracy and significantly reducing manual review processes by 90%. • Optimized PostgreSQL schemas and concurrent REST APIs, resulting in a 3× throughput improvement and 30% lower latency during peak load.

Education

Bachelor of Technology - BTech, Computer Software Engineering, Vellore Institute of Technology

Class 11-12, Birla bharati

Public links

github: https://github.com/Saunakghosh10

codeforces: https://codeforces.com/profile/saunakghosh10