Kunal Kaushal
AI Engineer
Building production RAG pipelines, multi-agent architectures, and real-time voice agents.
Currently AIML Trainee at Droisys.
kunal@portfolio:~
I'm an AI Engineer building production RAG pipelines, multi-agent systems, and real-time voice agents in Python. Currently at Droisys, developing an in-house recruiting platform combining hybrid search over 5,000+ resumes, an agentic workflow layer, and an AI phone interviewer.
Pursuing B.Tech in AI & Machine Learning (2023 - 2027) at GL Bajaj Institute of Technology and Management, focusing on high-performance retrieval architectures, LLMs, and distributed AI agents.
Where I've Worked.
AIML Trainee @ Droisys
Noida, India
Building Recruiter AI, an in-house platform automating candidate sourcing, interviewing, outreach, and resume processing.
- Built an agentic layer on Google ADK that lets recruiters create job descriptions, search candidates, and send outreach emails through natural-language prompts, automating ~80% of previously manual recruiting tasks.
- Developed a real-time AI voice interviewer that has conducted 50+ phone interviews end to end: Twilio telephony, STT/TTS, and voice activity detection streamed through Pipecat on a FastAPI backend, with Google Gemini driving conversations.
- Built a hybrid dense-sparse retrieval pipeline on OpenSearch, fusing HNSW vector search with keyword retrieval via Reciprocal Rank Fusion, powering semantic candidate search across 5,000+ resumes.
- Automated conversion of incoming resumes into company-standard templates, eliminating manual reformatting for recruiters.
AIML Intern @ Droisys
Noida, India
Engineered core RAG components, document extraction pipelines, and evaluated retrieval precision using RAGAS.
Archon: AI Knowledge Assistant
A high-performance RAG assistant for natural-language Q&A over 1,000+ technical documents. Designed hybrid retrieval combining FAISS dense search with BM25 keyword search, improving relevance on both exact-term and semantic queries, with answers grounded using Gemini 2.5 Flash.
- Python
- Gemini 2.5 Flash
- LangChain
- FAISS
- BM25
Selected Work.
Production systems built around retrieval, agents, and cloud-deployed inference.
Skills.
Grouped by capability - the systems and tooling I reach for in production.
AI & Agents
- RAG
- LLMs
- AI Agents
- Multi-Agent Systems
- Embeddings
- Ollama
- MCP
Retrieval & Evaluation
- Hybrid Retrieval
- HNSW
- BM25
- Reciprocal Rank Fusion (RRF)
- RAGAS
Frameworks & Backend
- Google ADK
- LangChain
- LangGraph
- Pipecat
- FastAPI
- Flask
- REST APIs
- Pydantic
Data & Vector DBs
- OpenSearch
- FAISS
- ChromaDB
- Pinecone
- pgvector
- SQL
- NoSQL
Cloud & DevOps
- GCP (Cloud Run)
- AWS (EC2, S3, IAM)
- Docker
Programming & CS
- Python
- C/C++
- Data Structures & Algorithms
- Object Oriented Programming
Developer Tools
- Git
- GitHub
- Bitbucket
- Jira
- Postman
- VS Code
- GitHub Copilot
- Cursor
- Claude Code
- Codex
- OpenCode
- Antigravity
Academic Journey & Certifications.
B.Tech in Artificial Intelligence & Machine Learning
GL Bajaj Institute of Technology and Management, Greater Noida
12th Grade (Senior Secondary)
Delhi World Public School Noida Extension
10th Grade (Secondary)
Delhi World Public School Noida Extension
Certifications
AWS Cloud Foundations
Amazon Web Services
AI Engineer Core Track: LLM Engineering, RAG, QLoRA, Agents
Udemy