087_Tarun

As an AI Engineer with over four years of experience, I am driven by a deep curiosity to explore complex challenges and a relentless ambition to achieve mastery in my field. I thrive on autonomous problem-solving, consistently delivering accountable and logical solutions that push the boundaries of what's possible.

Experience

H

AI Engineer

Hiring Plug · Feb 2026 – Present

Not yet confirmed

• Architected a production FastAPI backend (~88 Python modules, 10+ API surfaces) powering HiringPlug's AI-driven recruiting platform, serving 200K+ candidate profiles, 50K+ job postings, and 100+ agencies.
• Migrated HiringPlug's AI HR assistant from a single-intent MongoDB Atlas RAG system locked to one query type per conversation, degrading into hallucination after 5-6 turns, no pagination/deduplication, ~55s response time to a multi-agent architecture on Qdrant with a query orchestrator that decomposes requests across job, candidate, and FAQ agents within a single conversation, eliminating hallucination, adding proper pagination, and cutting response time to under 20s (as low as 15s); also introduced candidate summary generation during ingestion, previously absent, for complete and consistent retrieval metadata.
• Sole engineer building CoffeeCrew, an internal AI recruiting platform serving 2-3M candidate profiles: designed a multi-agent assistant where a task-understanding agent routes queries to specialist agents (JD analysis, candidate matching), with prompt-injection detection and content moderation on all pasted/uploaded JDs, achieving ~90% accuracy against a test set of injection/inappropriate-content attempts.
• Engineered semantic caching for common queries and model routing (lightweight tasks to Claude Haiku, complex tasks to larger models) with automatic multi-provider failover (Anthropic → OpenAI → Gemini) for cost efficiency and uptime.
• Built a parallel batch-scoring service processing candidates in batches of 25 against Qdrant-backed embeddings in 15-30s per batch, generating an LLM-based fit score and reasoning per candidate for filtering; also built JD and candidate parsing services (~3s per resume).
• Built an AI outreach system conducting candidate communication across WhatsApp, email, and LinkedIn in channel-appropriate tone, achieving ~90% accuracy in understanding and correctly shortlisting candidate replies, managing the funnel through to interview scheduling; also built an AI meeting-transcript scoring service.
• Owned CI/CD (GitHub Actions → Docker → AWS ECR → EC2), full test coverage, and platform-wide RBAC/rate limiting across all CC APIs; built and deployed a Model Context Protocol (MCP) server exposing recruiter chat as a tool, integrated live with ChatGPT, Claude, and Cursor.

AI Engineer

hiringplug™ · Feb 2026 – Present

Not yet confirmed

• Built HiringPlug's AI HR assistant API, serving 100K+ candidate profiles, jobs, and FAQs via multi-tenant RAG on MongoDB Atlas Vector Search; filtering and reranking improved top-5 retrieval precision by 20%, and grounded retrieval with guardrails reduced hallucinations in production answers.
• Built CoffeeCrew's AI recruiter assistant using Qdrant for candidate embeddings and vector search across conversational search, matching, and retrieval, backed by an async RabbitMQ ingestion pipeline (dead-letter queues, exponential retry) and deployed on AWS LightSail (Docker, CI/CD, Redis) as sole DevOps owner.
• Re-architected a single-intent RAG pipeline into a LangGraph multi-agent system for conversational candidate search, screening, JD intake, and meeting analysis, removing intent-lock so one conversation handles changing recruiter intent; reduced end-to-end query latency from 45s to 15s (~65%) via async workflows and streaming.
• Designed a streaming FastAPI backend for real-time token generation and persistent multi-turn recruiter conversations across chat, WhatsApp, email, and LinkedIn.
• Built recruiter search orchestration: natural-language profile building, specialist agents (skills, title, location, company), JD parsing (PDF/DOCX), and people-search query construction.
• Implemented LLM-based candidate matching/scoring with a 15% CTC/notice fit buffer to avoid false-rejecting borderline candidates, plus meeting-transcript analysis and safety guardrails.
• Built an MCP service exposing CoffeeCrew recruiter chat as a tool for agents/IDEs, enabling authenticated multi-turn search with full conversation state.
• Evaluated agent workflows on Azure AI Foundry using Microsoft Agent Framework alongside production LangGraph services, and implemented shared prompt-injection detection and content moderation across all LLM-facing flows.

AI Engineer – Generative AI / Agentic AI

Hexaware Technologies · May 2024 – Jan 2026

Not yet confirmed

• Spearheaded a document similarity and matching system using GPT-40 and OpenAI embeddings with cosine similarity to automate matching between job descriptions and consultant profiles, replacing an earlier Sentence Transformer + Groq pipeline for improved speed and accuracy; prototyped retrieval with FAISS and Pinecone before selecting Qdrant for native persistence and cloud-hosted scalability.
• Directed a cross-functional team of 4 to architect a multi-agent system (Comparison, Ranking, Communication Agents), replacing manual recruiter comparison with parallel AI agent execution and real-time progress tracking.
• Developed a React.js + Tailwind CSS dashboard with live status updates, JD comparison, top-3 candidate matches, and an admin console for agent queue monitoring, supporting 50+ recruiters.
• Built FastAPI backend microservices with JWT-authenticated access and asynchronous request handling, cutting response time to ~5 seconds; cursor-based pagination reduced candidate list load times and user drop-off during browsing.
• Engineered a LangGraph multi-agent L1 automation system for ServiceNow, replacing multi-day manual back-and-forth ticket resolution with automated AI-driven triage, classification, and response across 500+ tickets/month; included a KB search and auto-resolve agent for routine tickets, with adaptive LLM clarification loops improving first-pass resolution.
• Refactored legacy code into modular components with PyTest unit/integration testing, raising test coverage from 60% to 85% and reducing recurring post-deployment bugs.

AI Engineer

Hexaware Technologies · May 2024 – Jan 2026

Not yet confirmed

• Spearheaded development of a document similarity system using GPT-4o, OpenAI embeddings, and cosine similarity to automate matching between job descriptions and consultant profiles.
• Led a team of 4 in designing and implementing a multi-agent architecture comprising Comparison, Ranking, and Communication Agents for modular execution and real-time progress tracking.
• Developed FastAPI-based backend microservices for semantic embedding generation, comparison logic, and JWT-authenticated user access, while building RESTful APIs and microservices with asynchronous request handling that improved response time by 80%.
• Implemented real-time AI workflow tracking using WebSockets, reflecting stages such as embedding generation, profile ranking, and email delivery.
• Developed a React.js + Tailwind CSS dashboard with live status updates, JD comparison, top-3 candidate matches, email tracking, and an admin console for agent queue monitoring and performance reporting.
• Optimized backend performance through cursor-based pagination, reducing load times by 65% and lowering drop-off rates by 30%.
• Refactored legacy code into modular components and implemented unit/integration testing with PyTest, increasing test coverage from 60% to 85% and reducing post-deployment bugs by 50%.
• Integrated automated email notifications and MySQL persistence for recruiter workflows, matching results, job descriptions, and user data.

React Js Intern

Chance App · Aug 2022 – Sep 2022

Not yet confirmed

What drives their work

The Architect

Before acting, you envision the complete system and its underlying structure.

087_Tarun's superpower

You excel at designing and implementing robust, end-to-end AI systems that deliver significant functional and performance improvements.

How 087_Tarun works

You are best suited for teams that value end-to-end system design, performance optimization, and continuous technical growth, particularly in AI and backend development.