Skip to content
View YatharthLakhera's full-sized avatar

Block or report YatharthLakhera

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
YatharthLakhera/README.md

πŸ‘‹ Yatharth Lakhera | AI Engineer & Product Builder

Portfolio LinkedIn Email

3X Founding Engineer | Building AI-Powered Products That Scale

I don't just build AI systemsβ€”I build products that ship fast, scale efficiently, and deliver measurable business impact. My approach: ruthless optimization, data-driven features, and zero tolerance for bloat.

πŸ† Top Wins

  • Built Financial AI Agent for a Stealth Startup helping raise $100k+ in funding.
  • Built ETL Aggregation Pipeline of 100M+ Data Points while reducing cost by ~81% with read latency of P95 <200ms.
  • Increase Insurance/Compliance Documents processing speed by 80%(5hr β†’ 1hr) and >95% better classification.
  • Built Customer Call Automation handling 30k+ calls/day with pickup-rate increase 25-40%(high variation due to external factors)

πŸš€ Impact So Far

Current Project : Data Engineering & Visualization Project

  • Building real-time analytics dashboard (Apache Superset-style) with LLM-powered query optimization
  • Achieved P95 latency <200ms through lazy async caching and smart data sync
  • Empowering stakeholders with self-service data insights across multiple databases

Previous Highlights:

  • πŸ—οΈ Founding Member @ Amalgamic.io: Built MVP AI Agent for Finance with multi-platform integrations (stealth mode)
  • πŸ’° Cost Optimization at Scale: Slashed AWS costs by 40% and GCP costs by 80% at Ayu Health
  • ⚑ Performance Engineering: Increased database performance by 10X, deployment speed by 2X
  • πŸ€– Automation Impact: Scaled customer support automation to reduce operational load by 50%+

πŸ’‘ What I Bring to the Table

As a Founding Engineer:

  • 0β†’1 product development with MVP-first mindset
  • Identify and eliminate bias-driven features; ship what moves metrics
  • Bridge business requirements and technical execution
  • Rapid prototyping with production-grade foundations

As an AI Engineer:

  • LLM application development (LangChain, AI Agents)
  • Smart cost/latency tradeoffs in production AI systems
  • Multi-modal AI integration (OCR, Vision, NLP)
  • Building feedback loops into AI products

As a Technical Leader:

  • System architecture for scale and reliability
  • Team leadership and technical mentorship
  • Infrastructure optimization obsessive
  • Data-driven decision making

πŸ› οΈ Tech Arsenal

AI/ML Stack:

  • Frameworks: LangChain, AI Agents, OpenAI, Anthropic
  • MLOps: Model optimization, cost management, monitoring

Backend & Infrastructure:

  • Languages: Python, Java (Dropwizard, Guice DI)
  • Cloud: AWS (Lambda, ELB, SQS, SNS, CloudWatch), GCP (Vision, OCR)
  • Databases: PostgreSQL, MySQL, Redis, Elasticsearch
  • Monitoring: New Relic, custom observability pipelines

Product Development:

  • No-Code/Low-Code: Bubble.io, FlutterFlow
  • API Design: RESTful, FastAPI patterns
  • DevOps: Docker, CI/CD pipelines, distributed systems (Redisson)

πŸ“Š By The Numbers

Metric Achievement
Cloud Cost Reduction 40% AWS, 80% GCP
DB Performance Gain 10X improvement
Automation Scale 50%+ operational load reduction
Deployment Speed 2X faster pipelines
Analytics Latency P95 <200ms
Founding Experience 3X startups (Amalgamic, launch.today, Ayu Health)

🎯 Case Studies

1. Real-Time Analytics with LLM Query Builder Challenge: Internal teams needed self-service analytics without SQL expertise Solution: Built dashboard with LLM-powered query generation, multi-DB support, and intelligent caching Impact: Sub-200ms P95 latency, empowered non-technical stakeholders

2. Infrastructure Cost Optimization (Ayu Health) Challenge: Runaway cloud costs impacting margins Solution: Systematically analyzed resource utilization, re-architected GCP Vision/OCR usage, optimized AWS infrastructure Impact: $XXX,XXX annual savings (40% AWS, 80% GCP reduction)

3. Scaling Automation for Growth (Ayu Health) Challenge: Customer support couldn't scale with user growth Solution: Built intelligent automation pipelines with WhatsApp chatbot, IVR integration, and workflow optimization Impact: 50%+ reduction in operational load, 25% call pickup improvement

πŸ† What Sets Me Apart

  • Bias-Free Feature Development: I kill features that don't move metrics, regardless of who suggests them
  • Speed + Quality: Shipped production systems in weeks, not months (revamped entire backend in 30 days)
  • Full-Stack Capability: From LLM agents to infrastructure, no-code tools to distributed systems
  • Technical Depth + Business Acumen: Every technical decision ties back to ROI and user impact
  • Founding Engineer DNA: Comfortable with ambiguity and rapid iteration while pushing product/growth stakeholders to validate before over-building
  • Optimization Mindset: Cost optimization is built into my architecture decisions from day one

πŸ“« πŸ’¬ Let's Connect

I'm always interested in:

  • AI/ML opportunities where I can leverage LLMs, RAG, and AI Agents
  • Founding engineer roles in early-stage startups building impactful products.

Currently?

  • πŸ’ͺ Building real-time data analytics platforms powered by LLM query optimization
  • πŸ“š Learning Agent Harness, Local LLM and Context Engineering

Building an AI-powered product? Need a founding engineer who ships fast and optimizes ruthlessly? Let's talk.


"I don't build features. I build systems that solve problems, backed by data, shipped with speed."

Pinned Loading

  1. claude-gdrive-mcp claude-gdrive-mcp Public

    Claude.ai-compatible Google Drive MCP server (Drive-only fork of taylorwilsdon/google_workspace_mcp with renamed/reshaped tools to match Claude.ai's built-in Drive connector contract)

    Python 2 1

  2. savepoint savepoint Public

    Savepoints for Claude Code sessions. Two slash commands to persist and resume session context across windows.

    4

  3. claudebook claudebook Public

    Layered, AI-friendly project documentation for Claude Code. Two slash commands to bootstrap and incrementally maintain CLAUDE.md and .claude/docs/.

    1

  4. HR-Assist-RAG HR-Assist-RAG Public

    This system help find best candidate profiles from a candidate Database

    Python

  5. CalendarBackendApp CalendarBackendApp Public

    This is an app which details with user creation, setting default/custom availability timings and booking events based on availability!

    Java