10/02/2026 updated
100 % available
AI Engineer | LLM & ML Systems | Production AI Products
Delhi, India
Worldwide
Btech in Artificial Intelligence And Data ScienceAbout me
Most AI tools fail because nobody opens them twice, that's the problem I work on. AI engineer at Joveo; three products shipped 0→1, now used across the company and by our enterprise clients.
GenAIPython (Programming Language)Backend
I build AI products end to end from the first conversation with a user through to production and adoption.
AI / LLM: RAG systems (pgvector, embeddings, chunking strategy), agent and MCP connector development, prompt design and evaluation, LLM cost and model-routing analysis across providers (Claude, Gemini, GPT), LangChain, LangGraph, fine-tuning. Built a RAG decision platform ingesting from Slack, email, call transcripts and reports, plus six agent connectors distributed company-wide.
Backend & infrastructure: Python, FastAPI, Flask, PostgreSQL (incl. pgvector), Redis, SQLAlchemy, Docker, Kubernetes, Terraform, Argo, Jenkins. Migrated a production service from Redis to PostgreSQL, own production deployments, and act as CODEOWNER on our main repo.
Observability & data: OpenTelemetry, Loki, Grafana, Snowflake, Looker. Instrumented org-wide LLM usage and used that telemetry to cut 24% of company-wide spend.
Auth & security: JWT, RBAC, Google and Microsoft/Entra SSO, email OTP, domain-level access control, API hardening (rate limiting, origin validation) ahead of security review.
ML: TensorFlow, Keras, Scikit-learn, XGBoost, LSTMs, Hugging Face.
Product: PRD and BRD authoring, user discovery, roadmap prioritization, vendor POCs and commercial evaluation, launch and enablement, adoption analytics, executive stakeholder management.
Experience: AI Engineer at Joveo - joined as an intern, converted to full-time in nine months, now own three products used across the company and by enterprise clients. Previously AI engineering internships at Honestly and Adopt.ai,
Experience
Professional experience0 - 5 years
Freelancer since2026
Languages
EnglishFluentHindiNative speaker
Project history
Scaled Beacon, an internal dashboard platform, to 400+ users including 190 external client users across 30+ enterprise clients. Architected Google, Microsoft/Entra SSO, and email OTP authentication. Cut 24% of company-wide Claude spend (~$6,300/year) with a model-routing analysis adopted as org-wide policy. Built and distributed 6+ internal AI agents and connectors pre-installed org-wide.
Replaced 4 hours of manual risk analysis with a 10-minute automated workflow across 51 client accounts by building Weatherman, an AI-driven RevOps intelligence tool. Flagged at-risk client revenue through Prompter, a RAG-based decision platform aggregating 8,600+ signals across 95 customers. Reduced CS/Support workflow time from 3 hours to 30 minutes (85%). Drove company-wide AI adoption as a founding member of Joveo's AI Center of Excellence.
Drove a 40% increase in new users within weeks and cut growth-team response time by 95% with a Twitter bot (Flask + Twitter API on Azure VM) automating real-time replies to mentions. Lifted brand-detection precision from 76% to 94% and cut misclassifications by 30% by augmenting YouTube transcripts with video descriptions.