AI Engineer with 6+ years in AI Agents, Distributed Systems & LLM
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Software Engineer with 6+ years of experience building high-throughput distributed systems and production-grade AI agents. Currently developing AI agent execution components with MCP-based tool orchestration, multi-agent workflows, and persistent memory systems for an automation platform exposing 580+ LLM-consumable integration endpoints. Previously engineered real-time data pipelines processing 2B+ events/day and high-concurrency messaging infrastructure handling 500M+ messages/day at 15K TPS. Deep expertise in event-driven architecture (Kafka, Redis, BullMQ), observability (OpenTelemetry, Prometheus, Grafana), and building reliable AI systems that bridge LLM reasoning with backend execution.
Indian Institute of Technology (IIT), Roorkee
Bachelor of Technology · Electronics and Communication Engineering
August 1, 2015 – June 30, 2019
Andai Platforms Pvt. Ltd.
Full Stack Developer
February 1, 2026 – Present
Remote, India
Sunserg Technologies Private Limited
Software Engineer
November 1, 2022 – November 30, 2025
Gurgaon, Haryana, India
Jio Platforms Limited
Software Engineer
July 1, 2019 – August 31, 2022
Navi Mumbai, Maharashtra, India
FLO – Workflow Automation Platform
February 1, 2026 – Present
Implemented persistent agent memory using a dual-layer architecture (Redis-backed sliding window for short-term conversation context and PostgreSQL pgvector for long-term semantic retrieval via cosine similarity, with token budget management). Collaborated on the unified tool registry powering the AI agent execution engine, integrating 3 tool types (platform nodes, flow-as-tool, external MCP servers via Streamable HTTP/SSE) to enable autonomous multi-step task execution. Contributed to the platform-wide MCP server exposing 580+ integration nodes as LLM-consumable tool endpoints for Claude Desktop, Cursor, and Windsurf clients. Contributed to the execution observability layer using OpenTelemetry SDK with OTLP HTTP trace exporter to Jaeger, and authored Prometheus metrics for flow-level latency histograms (p50/p95/p99) and failure rate counters surfaced through Grafana dashboards. Contributed to hardening the job execution path with Redis-backed circuit breakers using atomic Lua scripts, BullMQ failed-jobs queue with custom DLQ routing and bull-board admin dashboard, and Redis SET NX idempotency guards – reducing silent job loss during partial failures.
SwitchKart – Re-Commerce Platform
November 1, 2025 – February 1, 2026
Contributed to the deterministic device grading and pricing engine supporting 13 grades across 4 device categories (Android, iOS, Fold/Flip, Note/Ultra), storing monetary values as integer paise to eliminate floating-point errors. Developed segments of the multi-stage order pipeline covering quote generation, Aadhaar-based KYC with OTP verification, and AWS S3 storage for device images and KYC documents with denormalized order snapshots for historical pricing integrity. Containerized services using Docker with images pushed to AWS ECR and configured CI/CD via Azure Pipelines with layer caching optimization.
DotMyStyle – Personalized Fashion Discovery & E-Commerce Platform
November 1, 2022 – November 30, 2025
Built 4 Kafka-based real-time inventory pipelines (reservation, release, deduction, replenishment) with dedicated consumer groups, dead-letter queue for failed events, and event-driven stock synchronization across a distributed multi-warehouse architecture, improving operational throughput by 10%. Designed multi-warehouse allocation engine with priority-based fulfillment and split-order logic, enabling automatic stock distribution across warehouses when a single facility cannot satisfy demand. Reduced out-of-stock incidents by 10% through proactive inventory balancing. Implemented atomic stock reservation using Mongoose optimistic concurrency control with 3-retry mechanism on version conflicts, paired with a Redis caching layer (configurable TTL, automatic invalidation on mutations), reducing inventory lookup latency by 20% and preventing overselling during concurrent peak traffic. Built low-stock alerting system with per-product, per-warehouse threshold monitoring and Kafka-published state-transition alerts. Integrated Prometheus gauges for real-time alert counts and Grafana dashboards for operational visibility. Developed comprehensive transaction logging system recording every inventory operation with before/after stock snapshots, operation duration tracking, and order-level audit trails - improving debugging efficiency and enabling root-cause analysis for stock discrepancies.
Jio TrueConnect DLT Platform – TRAI-Compliant Messaging Infrastructure
July 1, 2019 – August 31, 2022
Contributed to a high-concurrency SMPP proxy handling 500M+ messages/day, achieving 15K TPS using a centralized Redis-based atomic ID generator and asynchronous processing pipeline, while enforcing TRAI DLT compliance (sender ID, template, and consent validation) for Jio's commercial messaging traffic. Worked on the real-time analytics ingestion pipeline processing 2B events/day using Apache Kafka and Apache Druid, enabling SQL-based BI dashboards and reducing issue resolution time by 30%. Developed a custom automated regression testing framework for SMPP protocols, reducing deployment lead time by 80% while ensuring strict protocol compliance. Built and scaled REST API-based microservices for the client onboarding portal using MongoDB, ensuring atomic transactions and reducing customer onboarding time by 50%. Coordinated a 3-member team, increasing production issue resolution rate by 25%.
Cultural Fit Analysis
The candidate's project diversity, ranging from workflow automation and e-commerce to messaging infrastructure, indicates adaptability and a broad interest in different problem domains. Their current role and recent projects align well with the target role of an AI Engineer, showcasing a proactive move into AI-centric development. The breadth of technologies and architectural patterns they've worked with suggests a continuous learning mindset and a willingness to tackle new challenges, which is a strong cultural fit for an innovative AI team.
Soft Skills & Operational Fit
The candidate demonstrates strong problem-solving skills, evidenced by their contributions to complex system architectures and optimizations. Their experience in coordinating a team and improving issue resolution rates suggests good collaboration and leadership potential. The detailed project descriptions indicate a methodical approach to development and a focus on operational excellence, including observability and fault tolerance.