Headline

AI Engineer & Technical Lead @ Interview Kickstart | LLMs, RAG, AI Agents, GenAI, Voice AI | Backend: Python, Django, AWS Serverless

About

I build AI systems that pay for themselves — LLM pipelines, RAG, AI agents, and voice AI on AWS serverless, from POC to production at scale. Technical Lead of the Applied AI pod (4 engineers) at Interview Kickstart, a role I reached in under five years after joining as an intern in 2021. What I've shipped: → Central Intelligence Platform — LLM call intelligence I architected from a POC: HubSpot webhooks, AWS Lambda, speech-to-text, LLM scoring of every sales call, automated coaching reports. 700K+ calls/month at 99.5%+ pipeline success; $400K incremental revenue. → Agentic support chatbot — RAG + tool calling across ticketing, student-account, and scheduling systems; resolves tickets end-to-end without the ops team. Ticket volume down 35%. → AI voice agents — outbound calling (VAPI, Plivo, Twilio SIP) for webinar reminders and lead reactivation. 10,000+ calls/month with real-time HubSpot sync; show-rates up ~12%. → AI mock interviewer — strict persona prompting with guardrails; 500+ interviews, adopted as our hiring screen. The unglamorous foundations are mine too: company-wide DevOps (AWS, GitLab CI/CD, CloudWatch, cost governance), a ~35% cut in LLM spend via model routing across Claude, OpenAI, AssemblyAI and OpenRouter, and engineering hiring for the pod. I've been shipping LLM features since December 2022 — before it was a job title. Earlier: Klarna and Stripe payment systems, a referral engine that did $1M+ in year one, and platform internationalization that grew India sales 70%+. Open to: Technical Lead, AI Engineer, Senior Backend, and Forward Deployed Engineer roles — India or international. → madhav9620@gmail.com

Location

Bengaluru

AI stack

aistudio

anthropic apis

antigravity

claude code

claude

codex

Work experience

Technical Lead at Interview Kickstart, Mar 2026 - Present · 6 mos, Leading the Applied AI pod (4 engineers): sprint planning, architecture reviews, and hiring — every engineering hire goes through my technical round. • Scaled our AI call-auditing system into the Central Intelligence Platform: 700K+ calls/month analyzed at 99.5%+ pipeline success, $400K incremental revenue from coaching insights and lead-quality signals • Leading an agentic support chatbot (RAG + tool calling) that resolves student tickets without the ops team — ticket volume down 35% • Cut LLM/API spend ~35% (~$5K/month) via model routing and prompt optimization (Claude, OpenAI, AssemblyAI, OpenRouter) • Own company-wide DevOps: AWS infrastructure, GitLab CI/CD, monitoring (CloudWatch, Sentry), cloud cost governance • Driving platform epics: PII security, HubSpot/FreshDesk→Salesforce migration, Zapier→N8N migration (150+ workflows) • Standardized AI-assisted development with custom Cursor + Claude Code rules for the team

Senior Software Development Engineer at Interview Kickstart, Jul 2024 - Mar 2026 · 1 yr 9 mos, Bengaluru, Karnataka, India, Built the company's generative-AI product line from zero: • Architected Sales Copilot (now CIP): HubSpot webhook → Lambda → S3 → speech-to-text → LLM call scoring → automated weekly reports to sales leaders, with BigQuery analytics. Scaled across 4 sales segments; backfilled 1,100+ historical calls • Built an AI voice-calling bot (VAPI/Bolna/Plivo, Twilio SIP) for webinar reminders and lead reactivation — 10,000+ calls/month, real-time HubSpot sync • Built an AI Mock Interviewer with strict persona prompting and guardrails; extended into our hiring screen • Shipped an objection-handling engine mining historical call transcripts (Pydantic-validated LLM outputs) • Productionized lead-scoring ML models on Lambda + ECR with GitLab CI/CD, feeding HubSpot + BigQuery • Mentored a junior engineer to independent ownership of the India rollout

Software Development Engineer 2 at Interview Kickstart, Jul 2022 - Jul 2024 · 2 yrs 1 mo, Bengaluru, Karnataka, India, • Launched Klarna BNPL end-to-end + Stripe subscription lifecycle (installments, invoicing, dunning) • Built async queueing microservice for high-volume HubSpot↔platform API traffic; fixed sync failures blocking 100+ live enrollment deals • Founding backend engineer for MLCode (new ML interview-prep product); reviewed 15+ MRs, onboarded 2 engineers • Integrated OpenAI APIs for content generation in Dec 2022; shipped an FAQ chatbot in Apr 2023 • AWS hardening: secrets cleanup, root MFA, cost-anomaly detection

Trainee at CETPA Infotech Pvt. ltd., Jun 2019 - Jul 2019 · 2 mos, Noida, Uttar Pradesh, India, ▪️Machine Learning Model to predict Handwritten Digitwith Accuracy of 87% ▪️Fine-tune parameters of machine learning models to further enhance the accuracy ▪️Analysis of Twitter Dataset for the Election campaign of 2019

Education

Bachelor of Technology - BTech, Computer Science, DAV UNIVERSITY, JALANDHAR