Shantanu Sharma

Work Experience

Senior Software Engineer | Compass | New York, NY | Apr 2023 – Present

  • Spearheaded cross-functional development efforts with 8 engineers distributed across 4 global time zones to create scalable lead generation verticals, seamlessly integrating 100+ real-estate data sources into Compass' core CRM infrastructure and its downstream ML recommendation systems serving 30,000+ agents.
  • Designed and shipped the MCP-compliant contact orchestration layer powering Compass' CRM agent integrations, replacing ad-hoc per-vendor integrations with a single contract that handles creation, verification, and agent-matching for millions of contacts and feeds Likely-to-Win scoring downstream.
  • Scaled internal data infrastructure to power Compass' Likely-to-Win ML model, a personalized recommendation system combining property sale probability with agent win probability, driving a 39% increase in outreach on new recommendations and 16% on re-ranked existing ones (A/B tested), expanding into orphan matching, agent referrals, and seller-side lead routing
  • Catalyzed over $2M in quarterly GCI (Gross Commission Income), directly funding product, data, and engineering initiatives; contributing to Compass securing the #1 US brokerage ranking (RealTrends, 2022–2025) by closed sales volume.
Technologies: TypeScript, Node.js, Go, Kafka, PostgreSQL, React, Databricks, MCP, Datadog, Kubernetes, AWS

Software Engineer II | Compass | New York, NY | Jul 2021 – Mar 2023

  • Engineered the Compass Leads Program, an intelligent and fair routing mechanism for the company's critical revenue-generating leads funnel, matching clients to real estate agents and mortgage and lending services (title, escrow, and OriginPoint mortgage origination) and directly contributing to a multi-digit percentage increase in commissions revenue
  • Built the ingestion pipelines and underlying database that captured agent-property geo-matching signals, homeowner relationship strength, communication recency, and historical performance metrics; powering Compass' Likely-to-Sell ML recommendation model which contributed over $150MM per year in incremental revenue since 2021
Technologies: TypeScript, Node.js, Go, gRPC, PostgreSQL, React, Datadog, Kubernetes, AWS

Web Developer | WebMD | New York, NY | Sep 2019 – Jun 2021

  • Co-founded and developed WebMD's Helios platform, an enterprise framework on Vue.js, Node.js, NGINX, Docker, Kubernetes, GitLab CI, and the ELK stack for centralized observability.
  • Led migration of 95% of legacy consumer frontend and backend services to Helios, paired with Cloudflare CDN rollout and Prometheus, Grafana, and PageSpeed Insights monitoring, driving measurable gains in SEO, Core Web Vitals, and ad revenue.
Technologies: Vue.js, Node.js, TypeScript, Docker, Kubernetes, NGINX, Cloudflare CDN, Prometheus, Grafana, GitLab CI

Associate Web Developer | WebMD | New York, NY | Jun 2017 – Aug 2019

  • Sole developer and maintainer of WebMD's Healthcare AI Assistant on Google Assistant and Alexa, using conversational AI and NLU to answer 72% of healthcare voice queries about symptoms, medications, and treatment guidance for millions of users.
  • Independently proposed and built a backward-compatible testing framework that reduced WebMD's full test runtime from 14 hours to 1 hour by running Omniture analytics tests asynchronously across self-configured Selenium slaves.
  • Developed the Mucinex-sponsored health chatbot that generated a multi-million dollar revenue stream via a $2 CPC advertising model, serving thousands of daily health queries.
Technologies: Node.js, TypeScript, Google Dialogflow, Alexa Skills Kit, WebdriverIO, Selenium, Chrome DevTools Protocol, Jenkins, Adobe Analytics (Omniture)

Backend Web Developer | New Jersey Institute of Technology | Newark, NJ | Jan 2018 – May 2018

  • Contributed bug fixes and performance improvements to a P2P assignment-and-grading platform serving NJIT's CS cohort, working across the ExpressJS API layer, MySQL schema, Redis caching, and NGINX deployment.
  • Reduced response times on hot endpoints by 90% by introducing a Redis caching layer and an archival strategy that moved completed-semester data out of the live query path.
  • Instrumented the service with ElasticSearch, Fluentd, and Kibana structured logging, replacing scattered per-service log files with a queryable centralized stack.
Technologies: Node.js, ExpressJS, MySQL, Redis, ElasticSearch, Fluentd, Kibana, NGINX

Education

  • M.S. in Software Engineering | New Jersey Institute of Technology | GPA: 3.5/4 (42 credits) | May 2018
  • B.E. in Computer Science | Shri Ramdeobaba College of Engineering | May 2015

Git Activity

Loading activity…

Achievements

  • Sole recipient of the "Avengers team player" award at the Compass Annual All-Hands (2022).
  • Awarded "Engineer of the Year" at WebMD's Annual Tech All-Hands (2018).
  • Recognized as "Upcoming Leader of the Year" at WebMD's Annual Tech All-Hands (2019).

Projects

forgeseal, svidmint & assayward - Zero Trust Supply Chain (Go) | Mar – Jun 2026

  • forgeseal: Generates CycloneDX SBOMs from 13 lockfile formats with Sigstore keyless signing, SLSA v1 provenance, and OpenVEX/OSV.dev triage. Built for EU CRA compliance.
  • svidmint: SPIFFE workload identity for serverless/edge (Lambda, Cloudflare Workers, Deno, GitHub Actions) via platform-native attestation. Issues X.509 + JWT SVIDs.
  • assayward: Runtime trust gate evaluating those attestations and identities against declarative policy, emitting explainable allow/deny/audit decisions from one Go core (Wasm, CLI, K8s, edge).
Technologies: Go, Sigstore, SPIFFE, OpenVEX, CycloneDX, Wasm, OPA/Rego, X.509, JWT, OSV.dev

anyq - Universal Message Queue Library | Dec 2025 – Mar 2026

  • TypeScript and Go abstraction unifying Redis Streams, Kafka, SQS/SNS, Pub/Sub, RabbitMQ, NATS, and Azure Service Bus behind one interface, with DLQ and circuit breaker built in. Portable across cloud and self-hosted environments.
Technologies: TypeScript, Go, Node.js, Redis, Kafka, AWS SQS/SNS, Google Pub/Sub, RabbitMQ, NATS, Azure Service Bus

Watch And Learn (WAL) - AI+ML Video Processing Pipeline | Jul 2025 – Present

  • Multi-agent LLMOps pipeline on Kubernetes with LangGraph-orchestrated agent routing, tool-use, and retrieval, plus eBPF observability, HPA/VPA autoscaling, and active-active multi-region IaC.
Technologies: TypeScript, Golang, Python, LangGraph, Kubernetes, Kafka, vLLM, Ray, Pulumi, eBPF, AWS

Better Call Claude & Dear Claude - MCP Servers for Claude| Jan 2026

  • Built open-source MCP servers enabling bi-directional Claude Code communication via voice, SMS, and WhatsApp, plus webhook integrations with GitHub, Linear, Jira, GitLab, Notion, and Obsidian. Securely delivered the same agentic capabilities later popularized by OpenClaw, predating its viral release by weeks, entirely on Anthropic's official Claude Agent SDK without violating TOS.
Technologies: TypeScript, Bun, Hono, Claude Agent SDK, MCP SDK, Twilio/Telnyx APIs, OpenAI Whisper, Tailscale, Octokit

Unified Auth Gateway for Microservices | Aug - Sept 2025

  • Built serverless auth gateway on Cloudflare Workers with OAuth at the edge, using KV-cached sessions to reduce service onboarding time
Technologies: Cloudflare Workers, Hono, Better-Auth, TypeScript, OAuth, Cloudflare KV

Technical Skills

  • Languages & Frameworks: Python, Go, TypeScript, Node.js, Vue.js, React
  • Infra & Cloud: Kubernetes, Docker, AWS, Cloudflare Workers, Kafka, PostgreSQL, Redis
  • AI / ML Platform: LangGraph, MCP, vLLM, Ray, Databricks
  • Observability & Data: Datadog, Prometheus, Grafana, OpenSearch, eBPF