Lead Software EngineerAI/LLM & Agentic Systems6+ years

Harshit Meena

6+ years scaling distributed MERN / Node.js systems on AWS to 100K+ users — now driving production GenAI: LLM orchestration, agentic workflows, and RAG pipelines with a focus on evaluation, observability, and cost efficiency.

View resume Contact me

Email GitHub LinkedIn

9-Engineer Pod Lead @ Air India•100K+ users on MyAI•Multi-Agent RAG Orchestration

Current role:Air India

Location:Gurugram, India

DTU alumnus · 2016–2020

Delhi Technological University (DTU)

B.Tech. in Information Technology

CGPA: 7.7/10. Foundations in distributed computing, systems architecture, and algorithms.

Impact

Production metrics

CONCURRENT SCALE

100K+

PEAK

Users scaled (MyAI)

LEAD SCOPE & GOVERNANCEACTIVE LEAD

9ENG

Engineers in Pod

Engineers led

INFRA COST EFFICIENCY

~25%

REDUCTION

Shipping cost reduction

AI PIPELINE VOLUME

200K+

Course plans served

REAL-TIME DATA THROUGHPUT

1M+

Analytics data points

Perspective

Distributed Backbone to Autonomous Agent Execution

Lead Software Engineer with 6+ years scaling distributed MERN / Node.js systems on AWS to 100K+ users, now driving production GenAI initiatives — LLM orchestration, agentic workflows (LangChain/LangGraph), and RAG pipelines — with a focus on evaluation, observability, and cost efficiency.

Proven team leader shipping products end-to-end in Agile environments, from service boundaries and workflow orchestration through observability and mentorship.

PATTERNS: EVENT-DRIVENPARADIGM: AGENTIC MULTI-NODESCOPE: 100K+ USERS

Leadership

Active lead

Leading a team of 9 developers

Owning sprint planning, system design reviews, code reviews, and mentorship while shipping enterprise platforms end-to-end — plus prior mentorship of 3 junior developers at Edfora.

>Sprint planning, system design reviews, and code ownership

>Mentorship across a 9-engineer pod and prior junior cohort

>End-to-end delivery of enterprise platforms

Air India · 9-engineer pod2025 — PRESENT

Experience

Engineering & Systems Timeline

2020 — present

Air India

Current

Jan 2025 — Present • GURUGRAM, INDIA

Lead Software Engineer

>Led end-to-end architecture and delivery of MyAI, an enterprise workplace platform scaled from ~30K to ~100K users, owning service boundaries for finance records, leave workflows, document storage, org hierarchy, policy distribution, ticketing, and multi-level approvals from design through production rollout.
>Architected request/approval backend services with RBAC, JWT-authenticated APIs, audit trails, and deterministic state transitions, replacing fragmented manual routing with API-driven workflow orchestration used daily across teams.
>Traded short-term migration complexity for long-term reliability by re-architecting MyAI Social (feeds, engagement graph, media) onto S3/CDN offload, MongoDB read-replica routing, and Redis cache-aside ahead of the 3x user scale-up, sustaining stable feed-read latency on high-traffic timeline queries.
>Leading a team of 9 developers, owning sprint planning, system design reviews, code reviews, and mentorship.

Stack:Node.jsMongoDBRedisAWSS3CDNJWTRBAC

UniCommerce eSolutions Ltd

May 2024 — Jan 2025 • GURUGRAM, INDIA

Senior Software Engineer

>Built a shipping-provider aggregation service evaluating carrier rates/constraints across integrations to select optimal shipping options, reducing shipping costs by ~25% for 150+ businesses.
>Owned order-management backend workflows for cancellations and global order search using MongoDB indexing and Redis caching, cutting manual cancellation effort by ~30% and improving lookup speed by ~40%.
>Delivered a logistics control-tower service aggregating shipment status/events into operations analytics, streamlining weight-dispute workflows and reducing resolution time by ~25%.

Stack:Node.jsMongoDBRedisAPIs

Edfora Infotech Pvt. Ltd (FIITJEE)

Jun 2022 — Apr 2024 • DELHI, INDIA

Senior Software Engineer

>Designed and built an internal ticket-management system from scratch (Helpdesk-style) with RBAC, SLA-based escalation workflows, and audit trails, reducing issue-resolution time by 35%.
>Led development of MyClasses, an online class-scheduling platform for 50K+ users, integrating Zoom APIs and WebSocket-based live notifications for virtual learning.
>Engineered the Glorify Quotient analytics engine, a proprietary teacher-efficacy metric processing 1M+ data points via MongoDB aggregation pipelines and pre-computed materialized views.
>Architected a course-plan engine using SQS queues and Redis cache-aside caching, achieving ~500ms retrieval across 200K+ plans. Mentored 3 junior developers.

Stack:Node.jsMongoDBRedisAWS SQSWebSocketsZoom APIs

Wingify Softwares Pvt. Ltd

Aug 2020 — Jun 2022 • KOCHI, INDIA (REMOTE)

Software Developer (Frontend)

>Enhanced VWO's visual editor for layout/element/event/design manipulation, enabling analysts to run A/B personalisation with reduced engineering dependency.
>Implemented React.js/SPA support in the visual editor and testing product with rendering and load-path optimisations, improving page-load performance by ~40% and engagement by ~25%.

Stack:React.jsSPAPerformance

Projects

Applied AI Architectural Blueprints

3 case studies

Project 01·Agentic workflow

Agentic Profile Matching

Featured

Multi-step LLM agentic workflows for candidate matching with tool-calling, structured outputs, and ranked recommendations.

TOPOLOGY: ASYNCHRONOUS AGENTIC EVALUATION PIPELINELANGGRAPH • VECTOR STORE • EVALUATOR

[Candidate Input]Schema Verified

[Embedding Node]Dense Vector

[Vector Index]Top-k Retrieve

[Agentic Evaluator]Deterministic Graph

[Scored Matrix]Ranked output

STAGE 01

[Candidate Input]

Resume / Spec JSON

Schema Verified

▶flow

STAGE 02

[Embedding Node]

text-embedding pipeline

Dense Vector

▶flow

STAGE 03

[Vector Index]

HNSW nearest filter

Top-k Retrieve

▶flow

STAGE 04 :: LLM AGENT

[Agentic Evaluator]

LangGraph re-ranking chain

Deterministic Graph

▶flow

STAGE 05

[Scored Matrix]

JSON confidence matrix

Ranked output

Multi-step agentic workflow with retrieval, evaluation, and ranked structured output

Problem

Candidate matching needed multi-step reasoning — resume retrieval, tool use, and ranked recommendations — not a single prompt.

Architecture

Built LangGraph/LangChain agentic workflows with tool-calling agents, structured outputs, resume retrieval, and MCP tooling, exposed through a React dashboard.

Stack

LangGraphLangChainFastAPIChromaDBGroqMCPReact

Outcome

End-to-end agentic matching pipeline with ranked recommendations and dashboard visibility for operators.

Full case study →

Project 02·Agentic workflow

RAG-Based Profile Matching

Featured

RAG matching engine with chunking/embedding pipelines, hybrid skill filtering, and LLM-generated match reasoning.

TOPOLOGY: HYBRID RAG RETRIEVALFLOW: GROUNDED

Chunk & embed

Resume pipelines · sentence-transformers

Hybrid filter

Skills + semantic retrieval · ChromaDB

LLM reason

Match explanations over JDs · Groq

Chunking → hybrid retrieval → grounded LLM match reasoning

Problem

Job–resume matching required semantic retrieval and explainable LLM reasoning over job descriptions.

Architecture

Built resume chunking/embedding pipelines, hybrid skill filtering, and semantic ranking that produce LLM-generated match reasoning.

Stack

FastAPIChromaDBsentence-transformersGroqReactTypeScript

Outcome

RAG matching engine that surfaces ranked candidates with grounded match explanations.

Full case study →

Project 03·Knowledge platform

Resume Analyser

Case study

LLM tool-calling resume assistant for read/search/summarise flows over PDF/DOCX/TXT.

Problem

Resume files needed scoped read/search/summarise tooling without exposing unconstrained filesystem access.

Architecture

Built an LLM tool-calling assistant with scoped filesystem tools over PDF/DOCX/TXT and generated summary outputs.

Stack

PythonStreamlitGroqLlama 3.3pdfplumberpython-docx

Outcome

Interactive Streamlit assistant for resume read/search/summarise workflows.

Skills

Technical Stack & Architectural Arsenal

60 modules

AI / LLMCore focus

OpenAI & Anthropic (Claude) APIsLangChainLangGraphRAGHybrid Search (BM25 + kNN, RRF, Reranking)EmbeddingsAgentic AIFunction/Tool CallingMCPPrompt EngineeringLLM Evals & Observability (LangSmith)GuardrailsSemantic Caching

Languages04

JavaScript (ES6+)TypeScriptPythonSQL

Backend10

Node.jsExpress.jsRESTful APIsGraphQLWebSockets (Socket.IO)gRPCMicroservicesBullMQKafkaEvent-Driven Architecture

Frontend08

React.jsNext.js (SSR / ISR)AngularRedux ToolkitReact Query (TanStack)Tailwind CSSHTML5CSS3

Databases04

MongoDB (Atlas, Aggregation, Change Streams, Vector Search)PostgreSQL (pgvector)RedisElasticsearch

Cloud & DevOps08

AWS (EC2, S3, SQS, Lambda, ECS, CloudWatch)DockerKubernetesGitHub ActionsJenkinsCI/CDTerraformNginx

Observability & Testing07

PrometheusGrafanaOpenTelemetryJestSupertestPostmanSwagger/OpenAPI

Practices06

System Design (HLD/LLD)Distributed Systems (idempotency, DLQs, rate limiting, caching)TDDCode ReviewsAgile/ScrumMentorship

NowIn progress

What I'm Exploring & Prototyping

Production GenAI — LLM orchestration, LangGraph agentic workflows, RAG pipelines, and LLM evals/observability (LangSmith) with a focus on cost efficiency.

Experimental / lab work

EducationAug 2016 — Jun 2020

Delhi Technological University (DTU)

B.Tech. in Information Technology

Delhi · CGPA: 7.7/10

Aug 2016 — Jun 2020

Contact

Initiate Contact / Technical Inquiry

Open for inquiries

MessageReady

meenaharshit30@gmail.com

Phone hidden

Harshit Meena

Distributed Backbone to Autonomous Agent Execution

Leading a team of 9 developers

Air India

[+] View details

UniCommerce eSolutions Ltd

[+] View details

Edfora Infotech Pvt. Ltd (FIITJEE)

[+] View details

Wingify Softwares Pvt. Ltd

[+] View details

Agentic Profile Matching

RAG-Based Profile Matching

Chunk & embed

Hybrid filter

LLM reason

Resume Analyser

What I'm Exploring & Prototyping