{ "info": { "name": "Umang Bhalla", "location": "(@mumbai, india)[https://www.google.com/maps/place/Mumbai,+Maharashtra]", "profession": "systems engineer", "born": "spring 2001", "socials": { "github": "(@umgbhalla)[https://github.com/umgbhalla]", "x": "(@umgbhalla)[https://x.com/umgbhalla]", "discord": "(@umgbhalla)[https://discord.com/users/umgbhalla]", "linkedin": "(@umgbhalla)[https://linkedin.com/in/umgbhalla]" } }, "timezone": "gmt+5:30", "expertise": [ "harness engineering", "automation", "infrastructure", "system design", "mlops", "observability", "data warehousing" ], "work": [ { "year": "2026 april - present", "company": "zonko.ai", "role": "ai engineer", "location/hq": "(@mumbai, india)[https://www.google.com/maps/place/Mumbai,+Maharashtra]", "work": [ "agents, infrastructure, and robotics", "partnered with enterprise clients as an fde, translating business requirements, use cases, and product feedback into integration strategies while influencing the harbor and luffy product roadmap", "building luffy, an ai coworker for knowledge-work teams combining workspace context, skills management, reusable scripts, persistent execution flows, and harbor-powered tool integrations", "architected luffy's execution runtime on cloudflare workers and sandboxes for scalable request processing, isolated agent sessions, and compute-intensive or long-running workloads", "building harbor, an ai workbench and integration platform providing unified access to mcp servers, apis, and cli tools through a common plugin and execution framework", "designed harbor's codemode execution layer, introducing js/ts-inspired tool orchestration for composing, chaining, and reusing tools across complex multi-step agent workflows", "developed harbor's runtime and plugin architecture on cloudflare workers and ai gateway: tool discovery, oauth integrations, secrets management, policy enforcement, model routing, skills, and reusable workflow automation" ], "tech": [ ".*" ] }, { "year": "2025 oct - 2026 march", "company": "upsurge", "role": "systems engineer", "location/hq": "(@bangalore, india)[https://www.google.com/maps/place/Bangalore,+Karnataka]", "work": [ "new systems, infrastructure, web3, and robotics", "built and shipped (@bhindi.io)[https://bhindi.io], (@openbio.tech)[https://openbio.tech], (@moltx.io)[https://moltx.io], and (@mogra.xyz)[https://mogra.xyz]", "santra: ai-powered hardware device to talk to your agent on the go, with companion app, integrations, and long-term memory", "wrote firmware for xiao seed ble and optimized bluetooth comm for fast transcription on santra", "built infrastructure for solana-powered applications running agave validator", "self-hosted infrastructure for lido and hoodi nodes", "deployed ethereum reth full node with monitoring infrastructure" ], "tech": [ ".*" ] }, { "year": "2024 april - 2025 oct", "company": "newengen.com", "role": [ "sre 1", "sre 2" ], "location/hq": "remote/us", "work": [ "migrated infrastructure from gcp cloud run to kubernetes", "optimized infrastructure reducing compute cloud spend by 80%", "designed istio service mesh infrastructure for kubernetes native api gateway and ingress", "implemented custom api authentication layer with firebase integration across all endpoints", "automated ci/cd pipelines for 50+ microservices with gitops implementation", "designed and built internal tool and infrastructure for campaign, ads, and adset level taxonomy collection", "designed and implemented llm agent systems for marketing analytics, report generation, and trend analysis on 2+tb bigquery data", "built transformer-based recommendation system for taxonomy tool including training pipeline, model refresh cycles, and evaluation systems", "developed internal dev tools for model training and rollout, agent evaluations, and secret management" ], "tech": [ "kubernetes", "gcp", "istio", "firebase", "python", "bash", "prometheus", "grafana", "pytorch", "qdrant" ] }, { "year": "2023 july - 2024 march", "company": "reint.ai", "role": "mlops consultant", "location/hq": "remote/(@australia)[https://www.google.com/maps/place/Australia]", "work": [ "built mlops monitoring system with version control and metric logging for internal model training runs", "implemented high frequency monitoring and alerting system for vms and production apis", "designed custom training infrastructure utilizing multiple gpus, spot instances, and checkpoint resumption for long training runs", "conducted ablation studies on transformer models for time series forecasting", "built logging and statistics infrastructure with elastic, logstash, and kibana" ], "tech": [ "docker", "gcp", "python", "bash", "elastic", "logstash", "kibana", "pytorch", "autotrain", "clearml" ] }, { "year": "2022 may - 2023 june", "company": "bytelearn.com", "role": "devops consultant", "location/hq": "remote/(@india)[https://www.google.com/maps/place/India]", "work": [ "optimized aws infrastructure achieving 77% reduction in annual cloud spend", "migrated infrastructure to kubernetes", "migrated cloud infrastructure from aws to gcp", "integrated opentelemetry for distributed tracing and monitoring", "implemented high availability monitoring and observability stack with opentelemetry, promscale, tsdb, grafana, and prometheus", "deployed multi-node mysql and postgres clusters with perconadb", "managed ci/cd pipelines and uptime for 30+ services", "automated testing environment setup and deployment", "conducted security and compliance audits for gcp and kubernetes environments" ], "tech": [ "kubernetes", "argocd", "docker", "gcp", "python", "bash", "prometheus", "grafana", "mysql", "postgres", "perconadb", "redis", "tsdb" ] } ], "education": [ { "year": "2020 - 2024", "degree": "bachelor of Technology", "major": "information technology", "institution": "indian institute of information technology, bhopal", "location": "bhopal, madhya pradesh" } ], "format_source": "curl https://umgbhalla.xyz | jless" }