Bhavesh Saini

Software Engineer

Bengaluru, Karnataka, India1 yr 11 mos experience
AI EnabledAI ML Practitioner

Key Highlights

  • Delivered AI systems for Fortune 500 clients.
  • Achieved 98% improvement in document generation time.
  • Awarded multiple recognitions in under 18 months.
Stackforce AI infers this person is a Backend Engineer specializing in Fintech and Healthcare solutions.

Contact

Skills

Core Skills

PythonFastapiLangchainGenerative AiJavaReact.js

Other Skills

AWSSpring BootSSEAWS BedrockRedisasyncioPDF processingS3OpenAI APIsLLMPostgreSQLData StructuresProblem SolvingAmazon DynamoDBDatabase Management System (DBMS)

About

Backend Engineer | Python · FastAPI · LangChain · AWS Bedrock · SpringBoot | GenAI & Distributed Systems | HashedIn by Deloitte I build production-grade AI systems and high-throughput backend APIs that solve real problems at scale. 2 years at HashedIn by Deloitte, shipping concurrent distributed systems for Fortune 500 clients — Amgen and Santander Bank. ────────────────────────────── 🔑 What I've shipped: ▸ Streaming RAG API (Santander Bank) FastAPI + SSE on AWS Bedrock with Redis response caching, Cohere chunk reranking, and asyncio parallelization. Reduced end-to-end API latency from ~10s → under 4s with 20+ observability metrics tracked per request. ▸ LLM Answer Verification System (Santander Bank) 6-phase pipeline — embedding-based filtering, parallel PDF processing, concurrent LLM calls, PyMuPDF annotation, S3 signed URL delivery. Cut token usage by ~50%. ▸ Agentic AI Endpoint (Santander Bank) LangChain tool-calling agents autonomously orchestrating multi-tool reasoning chains streamed in real time via SSE. ▸ GenAI Regulatory Platform (Amgen) Automated pharmaceutical document generation. Turnaround: 8+ weeks → under 1 day (98% improvement). 90% section-level LLM accuracy on compliance workflows. ▸ Legacy Code Migration Tool (Internal) LLM-powered conversion of Oracle stored procedures + TIBCO workflows → Java Spring Boot microservices. 30% reduction in manual developer effort across a 5-engineer team. ────────────────────────────── 🏆 Recognition at HashedIn by Deloitte: 4× Award winner — Rising Star · Top Impactor · 2× Excellence — in under 18 months. Google Professional Cloud Architect certified (May 2025). ────────────────────────────── 💻 Tech Stack: Python · FastAPI · Java · Spring Boot · LangChain · LangGraph · asyncio · AWS (Bedrock, S3) · Redis · PostgreSQL · PGVector · MongoDB · Docker · OpenAI APIs · RAG · LLMs · SSE ────────────────────────────── 📩 Open to SDE / Backend Engineer / AI Engineer roles at product companies. Connect or message — happy to talk.

Experience

1 yr 11 mos
Total Experience
1 yr 11 mos
Average Tenure
1 yr 11 mos
Current Experience

Hashedin by deloitte

3 roles

Software Engineer - II

May 2026Present · 1 mo

PythonFastAPIAWSSpring BootLangChainGenerative AI

Software Engineer - I

Jul 2024May 2026 · 1 yr 10 mos

  • Built a real-time streaming RAG API (FastAPI + SSE) on AWS Bedrock Knowledge Bases with history-aware query reformulation, Cohere-based chunk reranking, and Redis response caching (SHA256-keyed, TTL-based); reduced end-to-end API latency from ~8-10s to under 4s by parallelizing retrieval via asyncio.gather and enabling Bedrock prompt caching, with a custom latency tracker measuring 20+ sub-step metrics.
  • Designed and owned a 6-phase LLM answer verification pipeline that filters relevant source documents via embedding-based KB search, downloads PDFs in parallel (ThreadPoolExecutor), runs concurrent LLM verification and citation calls, applies PyMuPDF highlights, and uploads annotated PDFs to S3 with signed URLs, cutting token usage by ~50%.
  • Developed an agentic AI streaming endpoint using LangChain tool-calling agents to autonomously orchestrate multi-tool reasoning chains (pricing analysis, CSV ingestion, knowledge base search) over complex user queries, streaming results incrementally via SSE in real time.
  • Built a GenAI platform automating pharmaceutical regulatory document generation for Amgen, reducing delivery turnaround from 8+ weeks to under 1 day (98% improvement in cycle time) by engineering S3-to-PostgreSQL ingestion pipelines that parsed 200+ unstructured PDF/Word documents daily with zero manual preprocessing.
  • Achieved 90% section-level LLM accuracy on pharmaceutical compliance documents via multi-step LangChain + OpenAI API workflows, reducing manual review cycles by 3x.
  • Engineered an LLM-powered code migration tool converting legacy Oracle stored procedures and TIBCO workflows to containerized Java Spring Boot microservices, reducing manual developer effort by 30% across a 5-engineer team.
LangChainGenerative AIFastAPISSEAWS BedrockRedis+4

Software Engineer Intern

Apr 2024Jul 2024 · 3 mos

  • Architected a full-stack Employee Management System using Spring Boot and ReactJS with JWT-secured REST APIs for frontend-backend communication
  • Built a RAG application using LangChain and OpenAI to enable intelligent querying across multi-format documents (PDF, DOCX, PPTX, XLSX) with Google Search API integration for real-time retrieval
LangChainReact.js

Salesforce

Virtual Salesforce Intern

Apr 2023May 2023 · 1 mo · Bengaluru, Karnataka, India · Remote

  • Completed Salesforce developer training covering CRM configuration, Apex programming, and Lightning components
  • Earned official Salesforce certification upon program completion

Education

Dayananda Sagar College of Engineering, BANGALORE

Bachelor of Engineering — Information Technology

Jan 2020Jan 2024

Stackforce found 100+ more professionals with Python & Fastapi

Explore similar profiles based on matching skills and experience