Vetted Vector Search Professionals

“Built and shipped a production LLM-powered customer support assistant using a Python/FastAPI backend with RAG (embeddings + vector search) over internal docs and product/operational data. Instrumented the system with logging/metrics and ran continuous eval loops; post-launch improvements focused on retrieval quality (chunking/ranking) and performance/cost tradeoffs (query classification, caching, validation guardrails).”

“Built and deployed a production RAG-based internal document Q&A system using LangChain, vector search, and a dockerized FastAPI LLM service. Focused on reliability by systematically reducing hallucinations and improving retrieval through prompt grounding/abstention strategies, chunking and top-k tuning, and iterative evaluation with logged metrics and manual validation.”

“Built a university campus assistant chatbot (BabyJ/WWJ) using RAG and agentic routing with a FastAPI + React stack and JWT auth, focusing heavily on production concerns like latency and reliability. Uses techniques like speculative prefetching, smart intent routing, and rigorous eval/testing (golden sets, regression, edge cases) while collaborating closely with campus admin/advising teams to iterate based on real user feedback.”

“Backend/ML engineering candidate focused on fintech automation who architected a zero-to-one agentic/LLM-enabled system to reconcile messy financial documents and bank transactions, reporting ~40% operational efficiency gains. Experienced migrating monoliths to event-driven microservices with incremental rollout via reverse proxy, and implementing production-grade security (OAuth2/JWT, RBAC, Supabase RLS) plus resilience patterns (timeouts/retries under concurrency).”

“Built an end-to-end car maintenance application from design through deployment using TypeScript/React with a Node.js + Firebase backend, including OBD integration. Uses mock servers and a layered frontend architecture (UI/state/services separation) to iterate quickly on UX without being blocked by backend or hardware dependencies.”

“Built and deployed a production bilingual (Bengali/English) AI virtual assistant that replaced IVR for telecom customer service at massive scale (~15M users), integrating ASR/TTS, Rasa dialogue management, and custom NLP. Overcame low-resource Bengali data and noisy call-center audio with synthetic data augmentation and transformer fine-tuning, achieving significant production gains including ~50% reduction in support calls.”

“Built and deployed an AI customer representative for iCore used at the IEE convention (2025), serving 100+ users in a day; implemented RAG with a vector database and scaled reliability via Docker and Google Cloud. Also has hands-on experience with multiple agent orchestration stacks (LangChain/LangGraph, Google AI Agent Development Kit, OpenAI SDK, Composio) and has delivered stakeholder-driven apps using prototyping and MVP scoping.”

“Built and shipped a production AI-driven privacy automation system that autonomously navigates data broker sites to submit opt-out/data deletion requests end-to-end, including robust CAPTCHA detection/solving (e.g., reCAPTCHA/hCaptcha/Cloudflare) via 2Captcha. Experienced in orchestrating stateful LLM agent workflows with LangGraph and hardening them for production with strict state management, retries/fallbacks, validation layers, and database-backed observability/audit logs, collaborating closely with legal/compliance stakeholders.”

“Built a production automated resume matching system using Python, FAISS vector search, and Selenium-based job scraping, including mitigation for IP blocking and heterogeneous site structures. Also develops LLM/RAG applications with LangChain, using Pydantic-guardrailed structured outputs and LLM-as-a-judge evaluation (including a project focused on tone/semantics for a 3D avatar’s emotional responses).”